ZEF
Zentrum für Entwicklungsforschung
Center for Development Research
University of Bonn
Working Paper 139
Doris Wiesmann, Hans Konrad Biesalski, Klaus von Grebmer, Jill Bernstein
Methodological Review and Revision of the
Global Hunger Index
ISSN 1864-6638
Bonn, October 2015
ZEF Working Paper Series, ISSN 1864-6638
Center for Development Research, University of Bonn
Editors: Joachim von Braun, Manfred Denich, Solvay Gerke, Anna-Katharina Hornidge and Conrad
Schetter
Authors’ addresses
Dr. Doris Wiesmann
Independent Consultant
Halendorf 3, 23744 Schönwalde, Germany
Tel. 0049 (0)4528-9135 176
E-mail: d.wiesmann@gmx.com
Prof. Dr. Hans Konrad Biesalski
Institute of Biological Chemistry and Nutritional Science (140), University of Hohenheim,
Garbenstr. 30, 70593 Stuttgart, Germany
Tel. 0049 (0)711-459 24112; Fax 0049 (0)711-459 23822
E-mail: hans-k.biesalski@uni-hohenheim.de
www.uni-hohenheim.de
Dr. Klaus von Grebmer
International Food Policy Research Institute (IFPRI),
2033 K St NW, Washington, DC 20006-1002, USA
Tel. 001 240-381 0611
E-mail: k.vongrebmer@cgiar.org
www.ifpri.org
Jill Bernstein
Independent Consultant
Tel. 001 321-693 3215
E-mail: jtwbernstein@yahoo.com
Methodological Review and Revision of the
Global Hunger Index
Doris Wiesmann, Hans Konrad Biesalski, Klaus von Grebmer, Jill Bernstein
Abstract
The Global Hunger Index (GHI) is a multidimensional measure of hunger that considers three
dimensions: (1) inadequate dietary energy supply, (2) child undernutrition, and (3) child mortality.
The initial version of the index included the following three, equally weighted, non-standardized (i.e.
unscaled) indicators that are expressed in percent: the proportion of the population that is calorie
deficient (FAO’s prevalence of undernourishment); the prevalence of underweight in children under
five; and the under-five mortality rate. Several decisions regarding the original formulation of the GHI
are reconsidered in light of recent discussions in the nutrition community and suggestions by other
researchers, namely the choice of the prevalence of child underweight for the child undernutrition
dimension, the use of the under-five mortality rate from all causes for the child mortality dimension,
and the decision not to standardize the component indicators prior to aggregation. Based on an
exploration of the literature, data availability and comparability across countries, and correlation
analyses with indicators of micronutrient deficiencies, the index is revised as follows: (1) The child
underweight indicator is replaced with child stunting and child wasting; (2) The weight of one third
for the child undernutrition dimension is shared equally between the two new indicators; and (3) The
component indicators of the index are standardized prior to aggregation, using fixed thresholds set
above the maximum values observed in the data set. The under-five mortality rate from all causes is
retained, because estimating under-five mortality attributable to nutritional deficiencies would be
very costly and make the production of the GHI dependent on statistics about cause-specific
mortality rates by country and year that are published irregularly, while the expected benefits are
limited.
Keywords: Global Hunger Index, composite index, food insecurity, child undernutrition, stunting,
wasting, underweight, child mortality, micronutrient deficiencies
JEL codes: I13, J13, O15
ii
Table of Contents
Abstract
ii
List of Tables
iv
List of Figures
v
Acknowledgements
vi
Acronyms and Abbreviations
vii
1
Introduction
1
2
The child undernutrition dimension of the Global Hunger Index
3
3
The child mortality dimension of the Global Hunger Index
9
4
Standardization of component indicators
14
5
Conclusion
20
References
23
Appendix A: Nutrition indicators proposed for the Sustainable Development Goals (SDGs) and not
recommended for inclusion in the Global Hunger Index
27
Appendix B: Reasons to include child mortality in the Global Hunger Index
30
iii
List of Tables
Table 1: Suggested priority nutrition indicators for the post-2015 Sustainable Development Goals
3
Table 2: Spearman rank correlations for child malnutrition indicators
4
Table 3: Spearman rank correlations for measures of child malnutrition and indicators of
micronutrient status
7
Table 4: Spearman rank correlations for different versions of the Global Hunger Index and indicators
of micronutrient status
8
Table 5: Summary statistics for the component indicators of the GHI
16
Table 6: Spearman rank correlations for changes in the Global Hunger Index from 1995 to 2005
18
Table 7: Spearman rank correlations for different versions of the Global Hunger Index
18
Table B-1: Global deaths in children younger than five years attributable to nutritional disorders
31
iv
List of Figures
Figure 1: Under-five mortality rate plotted against infant mortality rate
9
Figure 2: Under-five mortality in Rwanda, 1954-2013
10
Figure 3: Under-five mortality in Haiti, 1956-2013
11
Figure 4: Overview of dimensions and indicators of the revised Global Hunger Index
22
Figure B-1: Correlations of the Global Hunger Index with measures of hidden hunger
32
v
Acknowledgements
This review is part of a process that the International Food Policy Research Institute (IFPRI) initiated
eight years after the first Global Hunger Index was published in 2006. The intent is to assess whether
the dimensions, indicators and computation methods used for determining the annual Global Hunger
Index are state-of-the-art for a composite index of hunger.
The authors are grateful to Joachim von Braun, Director of the Center for Development Research,
whose suggestions were vital for this project. We also thank him and his colleague Nicolas Gerber for
their presentations at a workshop in Bonn on January 23, 2014, that outlined horizons for future
developments. The review also benefitted from consultations with colleagues at IFPRI, namely Marie
Ruel, Director, Poverty, Health and Nutrition Division; Lawrence Haddad, Senior Research Fellow,
Poverty, Health and Nutrition Division; and Harold Alderman, Senior Research Fellow, Poverty,
Health and Nutrition Division, who gave constructive feedback to the draft of this paper.
We also thank our colleagues and participants at the workshop in Bonn in January 2014 from
Welthungerhilfe (Ulrich Post and Andrea Sonntag), Concern Worldwide (Connell Foley and Olive
Towey), and IFPRI (Gwendolyn Stansbury) for their helpful input, and we are grateful to ClausBernhard Pakleppa from partnership for development (p4d) for moderating the discussions.
This project was funded by IFPRI, Welthungerhilfe and Concern Worldwide.
vi
Acronyms and Abbreviations
CM
Proportion of children dying before the age of five years
CST
Prevalence of stunting in children younger than five years
CWA
Prevalence of wasting in children younger than five years
DALY
Disability-Adjusted Life Year
DHS
Demographic and Health Survey
FANTA
Food and Nutrition Technical Assistance Project
FAO
Food and Agriculture Organization of the United Nations
GHI
Global Hunger Index
GNR
Global Nutrition Report
IDS
Institute for Development Studies
IGME
United Nations Inter-agency Group for Child Mortality Estimation
IFAD
International Fund for Agricultural Development
IFPRI
International Food Policy Research Institute
IRD
Institut de Recherche pour le Développement
MDG
Millennium Development Goal
NIMS
Nutrition Impact Model Study
p4d
partnership for development
PAF
Population attributable fraction
PUN
Proportion of the population that is undernourished
SD
Standard deviation
SDG
Sustainable Development Goal
UN
United Nations
UN ACC/SCN
United Nations Administrative Committee on Coordination/Subcommittee on
Nutrition
UNDP
United Nations Development Programme
UNICEF
United Nations Children’s Fund
UNSCN
United Nations Standing Committee on Nutrition
WFP
World Food Programme
WHA
World Health Assembly
WHO
World Health Organization
vii
1 Introduction
The Global Hunger Index (GHI) is a multidimensional measure of hunger that was first released by the
International Food Policy Research Institute (IFPRI) and Welthungerhilfe in 2006 (Wiesmann,
Weingärtner, and Schöninger 2006).1 The GHI considers three dimensions: (1) inadequate dietary
energy supply, (2) child undernutrition, and (3) child mortality. From the time of its original release in
2006 through the ninth release in 2014, the index has included the following three, equally
weighted2, non-standardized (i.e. unscaled) indicators that are expressed in percent:
1. the proportion of the population that is calorie deficient (FAO’s prevalence of
undernourishment, reflecting only quantitative deficits according to their definition)
2. the prevalence of underweight in children under five (WHO and other sources)
3. the under-five mortality rate (IGME) (Wiesmann 2006).
While the weighting of the three index components had been supported by empirical findings, i.e. by
the results of a factor analysis (Wiesmann 2004; Wiesmann 2006; Wiesmann, von Braun, and
Feldbrügge 2000), it is important to note for this methodological review that constructing composite
indices is no hard science. There is no commonly accepted best practice to construct international
indices, and arbitrary decisions about the choice of indicators, the standardization of the index
components or omission of standardization, and the weighting and aggregation function are
unavoidable. Therefore, composite indices will always be open to critique and should be revised from
time to time to reflect new thinking and the availability of new indicators or additional data.
A set of requirements for international indicators and indices has been postulated in the literature
that may guide such revisions. Indicators and indices should be relevant for national and
international policy debates, for example (see Wiesmann 2006, Table 3 on p.60, and Wiesmann 2004,
Chapter 2.1, for a fuller discussion). In addition, the data should be nationally representative,
comparable across countries, based on state-of-the-art measurement methods, and stem from
credible sources. Recent data need to be available for a large number of countries, and appropriate
estimation techniques have to be developed to fill data gaps.
Certain decisions regarding the calculation of the GHI are reconsidered in the following paper in light
of recent discussions in the nutrition community and suggestions by other researchers (Martorell
2008; Masset 2011). This concerns in particular
1. the choice of the prevalence of child underweight for the child undernutrition dimension,
2. the use of the under-five mortality rate from all causes for the child mortality dimension, and
3. the decision not to standardize (or “scale”) the component indicators prior to aggregation.
While FAO’s measurement approach and data to ascertain chronic calorie deficiency have frequently
been criticized, exchanging this indicator in the inadequate dietary energy supply dimension is not
discussed due to a lack of suitable alternatives with sufficient data availability that could replace it.
While modifying the dimension of inadequate dietary energy supply to also reflect diet quality would
be desirable from a conceptual point of view, such an endeavour faces practical challenges with
regard to the availability and appropriateness of indicators and data (see also the comments on a
dietary diversity indicator for women in Appendix A).
The average percent of dietary energy from starchy staples, for example, which is also based on
FAO’s Food Balance Sheets, is considered a rough indicator of diet quality, has wide data availability,
and shows a closer relationship with child stunting than measures of national dietary energy
1 The GHI is based on research by Wiesmann (2004) and Wiesmann, von Braun, and Feldbrügge (2000).
2 A factor analysis suggested approximately equal weighting (Wiesmann 2004; Wiesmann 2006).
1
availability. The report on the State of Food Insecurity in the World 2013 finds a much stronger
association between percent of energy from starchy staples and stunting in preschool children (R2
linear=0.46, corresponding to a Pearson’s correlation coefficient of 0.68) than between the
proportion of undernourished and stunting (R2 linear=0.28, corresponding to a Pearson’s correlation
coefficient of 0.53) (FAO, IFAD, and WFP 2013, p.25f). Based on cross-country regressions, Smith and
Haddad (2015) estimate that the percent of dietary energy from non-staples (that is, the inverse of
dietary energy from starchy staples) has greater potential to reduce child stunting in the future than
access to sanitation, women’s education, access to safe water, gender equality, and national food
availability.
Yet, for the purpose of index construction, the percent of dietary energy from starchy staples is
fraught with two disadvantages:
1. In contrast to the other indicators included in the GHI, the percent of energy from starchy
staples has an optimum greater than zero and the principle “the less, the better” does not
apply throughout. Determining an optimum value up to which reductions of the percent of
starchy staples should lead to decreases in GHI scores could prove difficult because the
optimum value for good nutrition and health varies with the composition of the local diet; it
depends, among other things, on the micronutrient content of the primary starchy staples
and the type of animal source foods consumed.
2. Until now, the data have been released with considerable delay: In 2015, the latest available
data on dietary energy from starchy staples were from 2009-2011. Thus, they are lagging five
years behind the most recent data on undernourishment, which are for 2014-2016 and partly
based on extrapolations (FAO 2015; FAO, IFAD, and WFP 2015). The inclusion of these data in
the 2015 GHI would affect the currentness of the index.
Nevertheless, the availability of indicators and data that measure or at least proxy diet quality should
be closely monitored, and diet quality should be considered for future revisions of the GHI if the data
situation improves.
2
2 The child undernutrition dimension of the Global Hunger
Index
The prevalence of underweight in children under five was the only nutrition indicator included with
an explicit reduction target in the Millennium Development Goals (MDGs), see United Nations
(2008), and has been used for the child undernutrition dimension of the GHI.
Yet, underweight is no longer the preferred indicator for monitoring child undernutrition because of
one drawback: It can mask the co-occurrence of normal weight and even overweight with stunting,
and thereby give a false impression of progress in reducing child undernutrition (Martorell 2008). In
some settings, interventions to improve nutrition can have the undesired side-effect of increasing
obesity. If underweight was chosen for monitoring, its reduction would be seen as an indication of
success despite the possible increase in overweight, while stunting may remain unchanged (Martorell
2008).
The Lancet series on maternal and child undernutrition uses stunting as an indicator of child nutrition
and complements it with measures of weight relative to height, namely wasting and overweight
(Black et al. 2008). Stunting was also found to be the best predictor of the long-term consequences
of child undernutrition on human capital (Victora et al. 2008).
In 2012, the World Health Assembly (WHA) endorsed six Global Nutrition Targets to be reached by
2025 (WHO 2012). Instead of child underweight, other anthropometric indicators were chosen to
monitor progress in the fight against child undernutrition: child stunting, wasting, and low birth
weight. All six WHA indicators are among the priority nutrition indicators proposed for the
Sustainable Development Goals (SDGs) by the United Nations Standing Committee on Nutrition
(UNSCN) (Table 1).
Table 1: Suggested priority nutrition indicators for the post-2015 Sustainable Development
Goals
Source: UNSCN 2014
While underweight reflects both acute and chronic undernutrition, stunting indicates only chronic
undernutrition, and wasting only acute undernutrition. Table 2 illustrates that underweight has the
strongest associations with the other two indicators of child undernutrition in the correlation matrix,
but also the highest correlation with overweight as a measure of child overnutrition.
3
Table 2: Spearman rank correlations for child malnutrition indicators
Underweight Stunting Wasting Overweight
Underweight
1.00
Stunting
0.82
1.00
Wasting
0.83
0.54
1.00
Overweight
-0.65
-0.35
-0.45
1.00
Nationally representative survey data from the period 2009-2013 retrieved in May 2014 were used. A cutoff of
-2SD below the median was used to define underweight, stunting and wasting, and a cutoff of +2SD above the
median to define overweight. Correlation coefficients could be calculated for 65 countries with data for the
four indicators, and all were significant at p-value < 0.01. If data were used for the 75 countries countries for
which data on underweight, stunting and wasting were available, the correlation coefficients among the three
child undernutrition indicators would remain virtually identical.
Data sources: UNICEF/WHO/World Bank 2013; WHO 2014a; UNICEF 2014; MEASURE DHS 2014; India, Ministry
of Women and Child Development, and UNICEF, India 2014.
The change in thinking about the most suitable indicator(s) of child undernutrition suggests replacing
the underweight indicator in the GHI with stunting and wasting. There are several advantages of
replacing underweight with both wasting and stunting, as opposed to using only stunting as a
substitute:
1. Acute child undernutrition, which can be serious and life-threatening, is not neglected if the
focus of the child undernutrition dimension is not exclusively on stunting.
2. Wasting and stunting do not only differ conceptually, they also diverge empirically and show
a limited and somewhat variable overlap at the individual level: A recent analysis of stunting
and wasting in the same children in five countries with a high burden of undernutrition found
that 30 to 65% of wasted children were only wasted and not stunted, and, conversely, that
35 to 70% of wasted children were both wasted and stunted (IFPRI 2015). The two indicators
therefore cannot be treated as substitutes for each other.
3. Wasting reacts more quickly to variations in the determinants of child undernutrition than
stunting, thereby making the GHI more sensitive to changes over time (whereas stunting can
ensure a certain stability in the child undernutrition dimension, since this indicator is not
prone to erratic fluctuations).3
4. Stunting shows generally lower correlations with indicators of micronutrient deficiencies
than underweight—possibly because data which relate these child undernutrition indicators
to one or more specific micronutrient deficiencies are lacking4—and therefore stunting tends
to weaken the association of the GHI with measures of micronutrient deficiencies (“hidden
hunger”) if it replaces underweight. Yet, this tendency can be counteracted by using a
3
In other words, wasting can be considered a flow variable, whereas stunting is more a stock variable that
shows cumulative effects over time. Other composite indices have also aimed to keep the balance between
capturing the effects of past developments and being sensitive to recent changes, and have taken advantage of
combining flow and stock variables for that purpose. For example, the education dimension of UNDP’s Human
Development Index used to include the adult literacy rate (a stock variable that changes slowly over time)
together with the combined gross school enrolment rates (flow variables that react quickly to investments in
schooling), until these indicators were replaced by the mean years of schooling and expected years of schooling
in order to reflect past trends and current conditions (UNDP 2009; UNDP 2010).
4
It has to be emphasized that the picture presented in this correlation analysis is necessarily incomplete: For
most micronutrient deficiencies, data and sometimes even indicators are sorely lacking. For the few indicators
of micronutrient deficiencies where nationally representative survey data exist, they are only available for a
limited number of countries. Children with less severe subclinical deficiencies show no clinical signs such as
xerophthalmia or rickets so that the massive effects of micronutrient deficiencies on the immune system,
development and health may remain unnoticed.
4
combination of stunting and wasting to replace underweight instead of stunting alone
(Tables 3 and 4).5
5. A revised GHI that includes stunting and wasting is more consistent with the GHI that has
been used to date (Spearman rank correlation coefficient: 0.98) than a GHI where
underweight was replaced by stunting only (Spearman rank correlation coefficient: 0.94; see
also Table 7 in Section 4). When transitioning to the new version of the GHI, a greater
consistency with the previous GHI could help to avoid sudden jumps in countries’ ranking
position that might be difficult to explain.6
However, including wasting in addition to stunting also has potential disadvantages related to the
particularities of this indicator:
1. Because wasting reflects acute undernutrition, this indicator is more volatile than stunting,
and sudden events such as man-made or natural disasters can lead to quick changes in
wasting rates (WHO 2014b). The sometimes erratic fluctuations make it very difficult to
generate reliable estimates of wasting, and WHO (2014b) abstains from presenting trends at
the global level—let alone at the regional or country level—and presents only a snapshot of
the global wasting rate for 2012. IFPRI (2014) states that methods to generate reliable data
on wasting trends are yet to be developed. For 75 out of 120 countries for which the 2014
GHI was calculated, survey data for wasting were available for the reference period 20092013 as of 2014 (using the data sources from which data on underweight were compiled for
the 2014 GHI). If data availability is similar for the period 2010-2015, wasting rates may have
to be estimated for more than 40 countries to calculate the 2015 GHI, and potentially
unreliable estimates would influence the ranking of these countries on the index; this
problem cannot be avoided.7
2. Wasting responds more strongly to seasonal changes than stunting. It is therefore more
important for this indicator that surveys are conducted over an entire year, in a way that is
representative for all seasons, especially in countries that show high seasonality such as
some in West and Eastern Africa. This is not always done: Surveys that are nationally
representative with regard to the sampling often cover only a few months of the year. In
addition, limited representativeness over time may make it even more difficult to produce
reliable estimates in cross-country regressions.
Regressions using wasting as a dependent variable in the models developed to predict child
underweight showed promising results, although the goodness of fit was not as good as for stunting
(personal communication with Yisehac Yohannes, IFPRI). Producing reasonable estimates of wasting
5
This pattern is also reflected in the association of child malnutrition indicators with a dietary diversity
measure that was developed as a proxy indicator for the diet quality of complementary child feeding (WHO
2010a). Spearman rank correlation coefficients with low dietary diversity among young children amount to 0.80
for underweight, 0.81 for the combination of stunting and wasting, 0.67 for stunting, 0.65 for wasting,
and -0.44 for overweight, and are significant (p-value < 0.01). For this analysis, low dietary diversity was
defined as the proportion of children 6-23 months who consume fewer than four out of seven food groups
(grains, roots and tubers; legumes and nuts; dairy products; flesh foods; eggs; vitamin-A rich fruits and
vegetables; other fruits and vegetables); data were taken from WHO (2010b) and Kothari and Abderrahim
(2010).
6
For example, among the 75 countries for which the “old” GHI with underweight (excluding estimates from
regression models) as well as several versions of a possible “new” GHI were calculated, India ranked 57 on the
“old” GHI, and the same if the component indicators were standardized; 45 out of 75 on a new GHI with
stunting instead of underweight, and 44 out of 75 if the component indicators were standardized; and 52 out
of 75 on a new GHI with stunting and wasting instead of underweight, with standardized components.
7
Survey data for stunting were also only available for 75 out of 120 countries for 2009-2013 as of 2014,
however, producing reliable estimates for stunting is of less concern than doing so for wasting (WHO 2014b).
5
for countries with missing data appears to be possible. Based on the weights presented in Section 5,
wasting makes up only one sixth of the index if used together with stunting for the child
undernutrition dimension. Thus, the effect of estimates with limited reliability cannot be very large.
Consequently, underweight is replaced with a combination of stunting and wasting, although
empirical analysis shows that this change has little practical significance for the ranking of countries:
Different versions of the GHI, using either underweight or a combination of stunting and wasting for
the child undernutrition dimension, and using standardized or non-standardized components, are
highly redundant in terms of their rank correlations (Table 7 in Section 4). Thus, the virtue of
replacing underweight with wasting and stunting lies mainly in the different message sent to policymakers about the type of programs that should be implemented to improve child undernutrition.
Further data explorations should be undertaken to examine whether it would be advantageous to
use child undernutrition rates for selected age groups, for example, for the age group 0-2 years or
the age group 3-5 years, instead of the rate for all under-fives. Child stunting is thought to be
partially reversible up to the age of two years, but irreversible after the second year of life, when the
full extent of the damage becomes apparent. The results of such explorations could be considered in
future revisions of the index.
6
Table 3: Spearman rank correlations for measures of child malnutrition and indicators of micronutrient status 1
Prevalence among children under
five:
Underweight
2
Stunting + wasting, combined
Stunting
Wasting
Overweight
3
Low urinary
iodine
concentration
among school
children
(N=54)
0.28 **
0.27 **
0.23 *
0.25 *
(N=50)
-0.11
Median urinary
iodine
concentration
among school
children
(N=65)
-0.19
-0.17
-0.05
-0.31 **
(N=59)
0.04
Night
blindness
among
preschool
children
(N=24)
0.21
0.35 *
0.27
0.28
(N=21)
-0.01
Night
blindness
among
pregnant
women
(N=39)
0.50 ***
0.44 ***
0.28 *
0.43 ***
(N=38)
-0.48 ***
Low serum
retinol
among
preschool
children
(N=50)
0.39 ***
0.40 ***
0.43 ***
0.26 *
(N=43)
-0.12
Anemia
among
preschool
children
Anemia
among
pregnant
women
(N=77)
0.54 ***
0.49 ***
0.46 ***
0.41 ***
(N=74)
-0.32 ***
(N=55)
0.37 ***
0.39 ***
0.26 *
0.41 ***
(N=50)
-0.18
Sum of
correlation
4
coefficients
2.47
2.51
1.98
2.36
-1.27
1 Nationally representative survey data were used for indicators of micronutrient status. The latest available data were matched with the child malnutrition indicators using the
years of the surveys and the GHI reference periods. N indicates the number of countries for which the correlation coefficients could be computed. The highest correlation
coefficient in each column is emboldened.
2 Prior to adding the prevalence rates of stunting and wasting, they were standardized as follows, using thresholds of 70% for stunting and 30% for wasting: stunting (stand.) =
stunting/70*100; wasting (stand.) = wasting/30*100. The thresholds were set slightly above the maximum values observed in nationally representative data collected between
1988 and 2013 (26.0% for wasting and 68.2% for stunting, see Section 4 regarding the standardization of indicators).
3 The correlation coefficients for overweight were included as supplementary information, they are not essential for evaluating potential new child undernutrition indicators
for the GHI. Data on overweight are not available for all countries with data on underweight, stunting and wasting, which explains the lower sample sizes; the correlation
coefficients in the first four rows of the table were only calculated for countries for which data for all three indicators (underweight, stunting, and wasting) were available.
4 The sum of correlation coefficients provides only a rough indication of the strength of the association of anthropometric indicators with the various micronutrient
deficiencies, as the number and selection of countries for which the correlation coefficients could be computed are quite variable. The correlation coefficients for the median
urinary iodine concentration are negative for indicators of undernutrition and positive for overweight, and were multiplied with -1 before adding them up with the other
correlation coefficients. A higher median urinary iodine concentration indicates that the population is less deficient in iodine on average, whereas higher values for the other
indicators of micronutrient status indicate that deficiencies are more prevalent.
Definitions and data sources for indicators of micronutrient status: Anemia: Proportion of preschool-age children whose hemoglobin level is less than 110 grams per liter, and
proportion of pregnant women whose hemoglobin level is less than 110 grams per liter (World Bank 2014; MEASURE DHS 2014; de Benoist et al. 2008). Iodine: Proportion of
school-age children with urinary iodine concentration <100 microgram per liter, and median urinary iodine concentration of school-age children (Andersson, Karumbunathan,
and Zimmermann 2012). Vitamin A: Proportion of preschool-age children with night blindness, proportion of pregnant women with night blindness, and proportion of
preschool-age children whose serum retinol level is less than 0.70 micromole per liter (WHO 2009a).
* significant at p-value < 0.10, ** significant at p-value < 0.05, *** significant at p-value < 0.01
7
Table 4: Spearman rank correlations for different versions of the Global Hunger Index and indicators of micronutrient status 1
2
Global Hunger Index
Components not standardized,
including:
underweight
stunting +wasting
stunting
Components standardized,
including:
underweight
stunting +wasting
stunting
Low urinary
iodine
concentration
among school
children
(N=53)
Median urinary
iodine
concentration
among school
children
(N=64)
Night
blindness
among
preschool
children
(N=24)
Night
blindness
among
pregnant
women
(N=39)
Low serum
retinol
among
preschool
children
(N=50)
Anemia
among
preschool
children
Anemia
among
pregnant
women
(N=77)
(N=55)
Sum of
correlation
3
coefficients
0.25 *
0.24 *
0.24 *
-0.12
-0.10
-0.08
0.56 ***
0.57 ***
0.51 **
0.43 ***
0.33 **
0.29 *
0.48 ***
0.52 ***
0.51 ***
0.55 ***
0.54 ***
0.53 ***
0.36 ***
0.34 **
0.34 **
2.75
2.63
2.50
0.27 *
0.26 *
0.25 *
-0.14
-0.15
-0.09
0.59 ***
0.57 ***
0.55 ***
0.46 ***
0.41 ***
0.34 **
0.50 ***
0.53 ***
0.52 ***
0.62 ***
0.61 ***
0.58 ***
0.44 ***
0.44 ***
0.42 ***
3.01
2.96
2.74
1 Nationally representative survey data were used for indicators of micronutrient status. The latest available data were matched with the GHI and its components using the
year of the survey and the GHI reference periods. N indicates the number of countries for which the correlation coefficients could be computed. The highest correlation
coefficient(s) in each column is/are emboldened.
2 The different versions of the Global Hunger Index include FAO’s proportion of undernourished and the under-five mortality rate besides the listed child undernutrition
indicators. The three index dimensions are equally weighted, and stunting and wasting are weighted equally within the child undernutrition dimension if used in combination.
For the index versions with standardized components, the indicators were standardized using the following estimated maximum values prior to aggregation: 80% for
undernourishment, 65% for underweight, 70% for stunting, 30% for wasting, and 35% for under-five mortality (see Section 4 regarding the standardization of indicators). The
standardized values X (stand.) for each indicator were calculated as follows: X (stand.) = X/estimated maximum value * 100.
3 The sum of correlation coefficients provides only a rough indication of the strength of the association of the index versions with the various micronutrient deficiencies, as the
number and selection of countries for which the correlation coefficients could be computed are quite variable. The correlation coefficients for the median urinary iodine
concentration are negative and were multiplied with -1 before adding them up with the other correlation coefficients. A higher median urinary iodine concentration indicates
that the population is less deficient in iodine on average, whereas higher values for the other indicators of micronutrient status indicate that deficiencies are more prevalent.
Definitions and data sources for indicators of micronutrient status: See notes to Table 3.
* significant at p-value < 0.10, ** significant at p-value < 0.05, *** significant at p-value < 0.01
8
3 The child mortality dimension of the Global Hunger Index
It has been suggested to replace the under-five mortality rate with under-two mortality in order to
relate the GHI more meaningfully to the nutritionally critical first 1,000 days between conception and
the second birthday. Aside from the question of whether the 1,000 days-timeframe that is critical for
nutrition outcomes is also pivotal for child survival, implementing this replacement would meet some
serious challenges because under-two mortality is not a standard indicator and is not available from
official statistics.
It is doubtful whether it would be worthwhile to estimate under-two mortality rates from official
data, because the replacement would be of little relevance for the GHI ranking and trends over time.
This can be demonstrated from an analysis of under-five mortality and infant mortality rates. The
infant mortality rate is a standard indicator available from official statistics and indicates the
proportion of children dying before the age of 12 months. Using data for 2013 for 131 countries from
IGME (2014),8 the Spearman rank correlation coefficient for the infant mortality rate and the underfive mortality rate was very close to unity and highly significant, amounting to 0.9965. The close
association of the two indicators is also evident from the scatterplot (Figure 1).
Figure 1: Under-five mortality rate plotted against infant mortality rate
The mortality rates are expressed in their original scale, namely deaths per 1000 live births. Data for 2013 from
131 countries were considered.
Source: Authors’ own presentation based on data from IGME (2014)
Figure 1 also illustrates that most children who die before their fifth birthday die within the first year.
For the 131 countries considered, infant mortality contributes between 57% and 92% of overall
8
The 130 countries selected for calculating the 2014 GHI (von Grebmer et al. 2014), which include former
Sudan, were chosen for this analysis, but Sudan and South Sudan were treated as separate countries.
9
under-five mortality, with lower shares for countries with higher mortality rates. Liu et al (2015)
estimate that 44% of the 6.3 million children who died in their first 5 years of life died in the neonatal
period, that is, in the 28 days after birth. Therefore, the differences between under-five mortality
and under-two mortality rates can be expected to be fairly small.
Another suggestion to revise the indicator for the child mortality dimension of the GHI is
conceptually very appealing, namely the idea to use estimates of under-five mortality associated
with undernutrition instead of under-five mortality from all causes. Deaths from causes not related
to undernutrition are not suitable to measure the extent of hunger, or food and nutrition insecurity.
This becomes very clear in the event of man-made or natural disasters, such as the genocide in
Rwanda in 1994 and the earthquake in Haiti in 2010 (Figures 2 and 3).
Under-five mortality rate
Figure 2: Under-five mortality in Rwanda, 1954-2013
Years
The blue line depicts the under-five mortality estimates that are fitted to existing annual data from various
surveys and censuses (shown in other colors). Under-five mortality is expressed in deaths per 1000 live births.
The grey band indicates the confidence interval of the estimates.
Source: IGME (2014)
10
Under-five mortality rate
Figure 3: Under-five mortality in Haiti, 1956-2013
Years
The blue line depicts the under-five mortality estimates that are fitted to existing annual data from various
surveys and censuses (shown in other colors). Under-five mortality is expressed in deaths per 1000 live births.
The grey band indicates the confidence interval of the estimates.
Source: IGME (2014)
IGME (2014) estimates that under-five mortality more than doubled in Haiti between 2009 and 2010,
soaring from 8% to 17% because of the earthquake and its after-effects.9 Other things being equal,
this sharp rise in child mortality would raise the (non-standardized) GHI by 3 points, and only part of
this increase can be thought to reflect rising food and nutrition insecurity in the wake of the disaster.
Large-scale disasters such as the genocide in Rwanda and the 2010 earthquake in Haiti tend to
worsen food and nutrition insecurity and increase related deaths, but it can be assumed that a large
share of the children who died were killed immediately during these events and did not die from
undernutrition, or from the effects of the vicious cycle of inadequate food intake, weakened immune
systems, infectious diseases, and decreased appetite and nutrient absorption.
Black et al. (2013) estimate that globally, 45% of child deaths among children under five can be
attributed to undernutrition, namely to fetal growth restriction, stunting, wasting, underweight,
deficiencies of vitamin A and zinc, and suboptimum breastfeeding. While undernutrition is thus
deemed to be the direct or indirect cause of a large proportion of child deaths, more than half of all
children who died before age five died from causes not related to nutritional deficiencies. To
complicate matters, the proportion of child deaths attributable to undernutrition varies and cannot
be thought of as constant across countries and regions.
In principle, it should be possible to estimate the contribution of various forms of undernutrition to
mortality rates at the country-level. The Disability-Adjusted Life Years (DALYs) published by WHO for
9
This estimated increase seems to be based on assumptions about the impact of the earthquake and not on
actual data. The graph in Figure 3 does not show census or survey data that confirm the projected spike in
under-five mortality.
11
2004 incorporate country-level estimates of age-standardized mortality for all age groups from the
following nutritional deficiencies: protein-energy malnutrition, iodine deficiency, vitamin A
deficiency, and iron-deficiency anemia (WHO 2009b). Some of the factors considered by Black et al.
(2013) for their more recent global estimate of child deaths attributable to undernutrition, such as
fetal growth restriction, stunting, zinc deficiency, and suboptimal breastfeeding practices, were not
taken into account.
The DALYs from nutritional deficiencies should not be used to replace the under-five mortality rate in
the GHI for a variety of reasons,10 but they exemplify that it is possible to estimate deaths
attributable to nutritional deficiencies at the country level. Black et al. (2013) arrive at their global
estimate of child deaths attributable to undernutrition by computing child deaths for UN subregions.
This involves assessing the fractions of deaths from various causes (diarrhea, measles, pneumonia,
and other infections, excluding malaria) that are attributable to stunting, underweight, wasting, and
severe wasting—the so-called population attributable fractions (PAFs)—for each subregion. The PAFs
are multiplied with corresponding age-specific and cause-specific deaths to estimate the number of
deaths attributable to underweight, stunting, wasting and its subset of severe wasting.
More research into the literature and available data would be needed to understand the details of
the method and find out if and how it could be applied at the country level, or if PAFs for subregions
would be precise enough to derive country-level estimates. The required effort exceeds the scope of
this paper, but such a research project might be considered for the future. Before embarking on such
an endeavor, the following potential disadvantages of using estimates of child mortality attributable
to undernutrition instead of under-five child mortality rates from all causes should be taken into
account:
1. The child mortality data for the GHI would solely be based on own estimates and no longer
on official statistics; relying on official statistics has the advantage that people who have
questions about the data can be referred to the official sources.
2. The reliability of the estimates at the country level may be debatable, considering that
multiple assumptions have to be made to obtain them.
3. Data that are published irregularly may be required for the estimation process, such as
deaths and death rates from various infectious diseases by country and year.11
4. A lot of data work would be needed on an annual basis to arrive at the estimates, whereas
using the under-five mortality rates published by IGME is very simple and straightforward.
The costs for estimating child mortality from nutritional causes would most likely be
considerable, whereas the additional insights and benefits may very well be limited.
Considering the imperfections of the under-five mortality rate as an indicator of hunger and the level
of effort required to estimate the contribution of nutritional deficiencies to child mortality for the
GHI, it has been proposed to drop the child mortality dimension entirely from the index. Yet, this
option is rejected because (1) premature death is the most serious consequence of hunger, and
10
The value judgments to account for the time lived with disabilities and to aggregate losses from disability
with life years lost due to premature mortality are highly questionable (Anand and Hanson 1997). During a
presentation at the University of Kiel in the 1990s, Amartya Sen pointed out that minimizing DALYs lost in a
cost-effective manner could be equated with “saving the able-bodied”. In addition, some nutritional factors
that contribute to mortality have been neglected. It is also considered advisable to remain focussed on child
mortality for the GHI, which would not be given if the DALYs, which refer to all age groups, were used. So far,
the data for DALYs and death rates by detailed causes have been published irregularly, and it is not clear if and
when they will be updated. The data that are available at present are too outdated to be used for the 2015
GHI.
11
Black et al. (2010) published such estimates for 2008, Liu et al. (2012) provided an updated analysis for 2010
with time trends since 2000. Two years later, Liu et al. (2015) presented annual data for 2000-2013 and
projections to inform post-2015 priorities.
12
children are more vulnerable than adults; (2) the contribution of nutritional disorders to child
mortality is only partially captured by stunting and wasting; and (3) child mortality has higher
correlations overall with micronutrient deficiencies than the other components of the GHI and
thereby strengthens the ability of the index to reflect hidden hunger (see Appendix B for more
detailed explanations).
13
4 Standardization of component indicators
Masset (2011) commented on the GHI: “The choice of adding up the percentage values of each
dimension without prior scaling is also questionable as this ends up attributing lower values to
changes in mortality rates, whose percentage values are much lower than those of undernutrition
and food availability.” The standardization or scaling of index components is usually applied to
harmonize different measurement units (Szilágyi 2000). Ryten (2000) emphasises that for the
purpose of aggregation, the single indicators need to be expressed in a common metric, or need to
be standardized. If the component indicators are already expressed in a common metric,
standardization is optional and not mandatory. The Hidden Hunger Index for preschool-age children,
for example, aggregates the prevalence rates of stunting, anemia due to iron deficiency, and vitamin
A deficiency by simple averaging, without prior standardization of the indicators (Muthayya et al.
2013).
However, if the ranges between the maximum and minimum values of the single components differ
substantially, aggregation without preceding standardization is questionable. Despite equal
weighting, the indicator with the largest variance may dominate the overall index if no
standardization is applied, introducing a kind of “unintentional weighting” (Noorbakhsh 1998, p.593;
Wiesmann 2004). Such “unintentional weighting” can be avoided by standardizing the index
components. The following arguments speak in favor of standardizing the component indicators of
the GHI prior to aggregation:
1. The maximum values and standard deviations of the three index components—and of
wasting and stunting as suggested replacements for underweight—differ greatly and are
lowest for wasting and child mortality (Table 5). Standardization would give greater emphasis
to the child mortality dimension relative to the other two index dimensions, balance the
contribution of wasting and stunting to the child undernutrition dimension, and diminish the
relative impact of the contested undernourishment indicator on the variation of the index.
2. Different methods of standardization are discussed in the literature and were tested for the
GHI; the analyses showed that standardization had little effect on the ranking of countries12,
yet, the influence on changes of the GHI over time was non-negligible (Wiesmann 2006). Due
to its greater variability, changes in the undernourishment indicator tend to drive changes in
the index more than changes in the other two component indicators. This has also become
evident from the trend analyses for individual countries performed for Chapter 2 of the
Global Hunger Index reports (see the examples of Iraq and Swaziland in von Grebmer et al.
2014, p.17).
3. Because indicators of micronutrient deficiencies and low dietary diversity correlate more
strongly with under-five mortality and also—in most cases—more strongly with child
underweight than with undernourishment (see Figure B-1 in Appendix B), standardizing the
component indicators can be expected to make the GHI better suited to reflect hidden
hunger.
4. The lack of standardization of the component indicators of the GHI has frequently been
criticized by other researchers as a matter of principle (Masset et al. 2011; personal
communication with Shenggen Fan, IFPRI).
12
The rankings based on index versions with standardized components essentially contained the same
information as the ranking of the GHI without standardization. For the total sample, the rank correlation
coefficients were above 0.99, and fairly above 0.90 for the three subsamples with low, medium, and high GHI,
showing the high redundancy of the index versions with standardized components vis-à-vis the GHI without
standardization (Wiesmann 2006).
14
On the other hand, standardizing the index components involves arbitrary decisions and has several
disadvantages. If the actual minimum and maximum values observed in the data set are used, the
“goalposts” tend to shift every year due to changes in the data. This would affect the intertemporal
comparability of the GHI.13 To avoid problems with intertemporal comparability due to
standardization, estimated maximum values that are safely above the actual past and present
maximum values observed in the data set can be used. Yet, this approach has also its downsides:
1. It is not possible to select estimated maximum values without arbitrariness. While it is
possible to find estimated maximum values that are unlikely to be exceeded by any country
in the near future, they may not be suitable for subpopulations in extreme emergencies
when undernourishment, child malnutrition, and child mortality soar, and could pose
problems for subnational disaggregation in selected countries.14
2. It can be argued that the proportion of undernourished and the prevalence of underweight
or of stunting and wasting in children, which are more immediate measures of hunger and
undernutrition than under-five mortality, should have a comparatively larger impact on
changes in the index. Because the under-five mortality rate is a final outcome with other
causal factors than hunger and undernutrition, and only about half of all child deaths are
attributable to undernutrition, a smaller impact of this component on the GHI might be
considered acceptable (Wiesmann 2006). If it were possible to estimate child mortality from
nutritional deficiencies at the country level with sufficient reliability and reasonable effort,
this would make standardization more convincing, but this option is at present not within
reach.
3. Standardization of any type makes it a bit more complicated to trace back levels and trends
in GHI scores to the index components. This is irrelevant for analysts, but might matter for
the communication with a lay audience. For example, it is easy to communicate that about
66 percent of children were stunted in 1995 in Bangladesh, and easy to understand how this
affects the GHI without standardization. The meaning of the corresponding standardized
prevalence of stunting in children of about 94% (which would be included in the GHI based
on an estimated maximum value of 70 percent for the data set) is less obvious.
Despite certain drawbacks of standardization, it is considered advisable to standardize the
component indicators of the revised GHI to balance the contribution of wasting and stunting to the
child undernutrition dimension, decrease the influence of variations in the undernourishment
indicator on trends over time, and respond to the frequently voiced critique about the lack of
standardization by other researchers.15
13
Given the case that the maximum value for the prevalence of undernourishment would change from 68 to 75
percent from one time period to the next, a country with a constant prevalence of undernourishment of 50
percent and no changes in the other two index components would have a different GHI score and might be
classified differently, although nothing may have changed with regard to its food security and nutrition
situation.
14
The GHI was subnationally disaggregated for Ethiopia, India, and Nepal (Schmidt and Dorosh 2009; Menon,
Deolalikar, and Bhaskar 2009; WFP Nepal 2009).
15
How standardization balances the contribution of stunting and wasting to the child undernutrition dimension
of the GHI, and thereby to the overall index, can be nicely illustrated with reference to the cutoffs for public
health significance set for these indicators by WHO (1995). At the national level, prevalence rates of 15% or
more for child wasting, and of 40% or more for child stunting are regarded as severe public health problems.
Although both forms of child undernutrition reached very serious levels according to the WHO definition,
stunting would contribute 2.7 times as much to the child undernutrition dimension as wasting if a 40%
prevalence was simply lumped together with a 15% prevalence, without prior standardization and with equal
weighting. Standardization with the set maximum values of 70% for stunting and 30% for wasting (see below)
would largely redress this bias, resulting in standardized scores of 40%/70%=57% for stunting and of
15
Table 5: Summary statistics for the component indicators of the GHI
Indicator and
reference year(s)
N
Mean
Standard
deviation
Minimum
value
Maximum
value
Percentage of undernourished in the population, average for:
1990-1992
97
24.5
16.8
0.5
70.2
1994-1996
117
22.3
17.2
0.6
72.4
1999-2001
117
20.5
16.1
0.7
76.5
2004-2006
118
18.2
15.4
0.1
75.6
2011-2013
120
15.2
14.2
0.4
67.3
Prevalence of underweight in children under five, incl. regression estimates, from:
1988-1992
104
19.8
13.2
0.9
61.5
1993-1997
126
16.8
12.6
0.5
55.2
1998-2002
127
15.7
12.5
0.5
46.3
2003-2007
129
13.9
11.8
0.4
43.5
2009-2013
128
11.8
10.5
0.3
45.3
1
Prevalence of underweight in children under five , from:
1988-1992
48
19.5
13.7
2.5
61.5
1993-1997
82
18.1
12.5
0.5
55.2
1998-2002
91
17.5
11.9
0.7
46.3
2003-2007
100
14.6
11.7
0.6
43.5
2009-2013
75
14.8
10.6
1.1
45.3
1
Prevalence of stunting in children under five , from:
1988-1992
44
37.4
14.3
10.8
63.4
1993-1997
79
33.4
15.6
1.6
68.2
1998-2002
91
32.6
14.9
3.0
63.1
2003-2007
99
28.5
14.8
2.3
57.7
2009-2013
75
27.5
13.5
4.3
57.7
1
Prevalence of wasting in children under five , from:
1988-1992
43
6.8
4.8
0.6
20.3
1993-1997
77
8.4
5.4
0.5
22.4
1998-2002
91
7.8
4.9
0.5
19.4
2003-2007
97
7.4
5.1
0.3
26.0
2009-2013
75
6.9
4.8
0.6
21.5
Under-five mortality rate for:
1990
107
10.2
7.0
1.3
32.6
1995
127
8.6
6.6
1.0
27.9
2000
127
7.6
5.9
0.8
23.4
2005
128
6.3
5.1
0.7
21.6
2012
130
4.7
4.0
0.4
18.2
1 The number of observations is smaller for these data because regression estimates were not included.
N = number of observations. Data sources: See von Grebmer et al. 2014, p.40.
15%/30%=50% for wasting, and resulting in only a 1.1 times higher share of stunting relative to wasting in the
child undernutrition dimension.
16
Moreover, standardization could indeed be shown to increase the correlation coefficients for the GHI
and indicators of micronutrient deficiencies, yet, the effects were mostly very small (compare the
correlation coefficients for standardized and non-standardized GHI versions in Table 4).
Considering the maximum values of the component indicators observed in the data set for 19882013 (Table 5), the following thresholds were used to compute the standardized index versions
presented in this document:
Percentage of undernourished:
80%, exceeding the observed maximum of 76.5%, and also the maximum in the provisional,
unpublished country-level estimates obtained from FAO
Prevalence of underweight in children under five:
65%, exceeding the observed maximum of 61.5%; this threshold is used to calculate a
standardized version of the GHI with its usual component indicators for the sake of
comparison
Prevalence of stunting in children under five:
70%, exceeding the observed maximum of 68.2%
Prevalence of wasting in children under five:
30%, exceeding the observed maximum of 26.0%
Under-five mortality rate:
35%, exceeding the observed maximum of 32.6%.
Without a doubt, these thresholds could be set differently. It should be noted, though, that small
variations in the thresholds will have no tangible effects on the ranking or the changes of the GHI
over time. Two sets of alternative thresholds for standardization were tested: One using the
maximum values in the data set as indicated above, and one using thresholds set 10% above the
maximum values (that is, 84.15% for undernourishment, 67.65% for underweight, 75.02% for
stunting, 28.60% for wasting, and 35.86% for under-five mortality). Using data from 2009-2013, and
wasting and stunting for the child undernutrition dimension, the Spearman rank correlation
coefficients for the standardized GHI with the selected thresholds (see above), and the standardized
GHI versions with (1) the maximum values in the data set, and (2) the maximum values increased by
10%, amounted to 0.9991 each (see also Table 7 for the correlations between standardized and nonstandardized index versions).
Relative changes from 1995 to 2005 were calculated for the different GHI versions with stunting and
wasting, namely for the non-standardized index and the three standardized versions—the selected
thresholds, the actual maximum values, and the maximum values augmented by 10%. Table 6
confirms the very high redundancy of relative changes in the three standardized versions: If they
were not shown with four decimals, the Spearman rank correlation coefficients would have to be
displayed as 1.000 because they are so close to unity.
Estimated minimum values can also be used for standardization. Yet, this is irrelevant for the data
used for the GHI, because the observed minimum values are already close to zero (Table 5). Only
stunting has a slightly higher observed minimum value greater than 1%. Using a threshold of 1% as
the estimated minimum value for standardizing an indicator with a maximum value of about 70%
would be meaningless in terms of its effects on the index. Moreover, the actual minimum values for
stunting and other component indicators of the GHI are likely to decrease in the future.
17
Table 6: Spearman rank correlations for changes in the Global Hunger Index 1 from 1995 to
2005
Changes from 1995 to 2005 (in
percent) in the Global Hunger
Index with:
non-stand. components
selected
thresholds
stand.
maximum
2
components
values
maximum
values + 10%
2
standardized components
selected
maximum
maximum
thresholds
values
values + 10%
non-stand.
components
1.0000
0.9460
1.0000
0.9343
0.9985
1.0000
0.9356
0.9985
0.9998
1.0000
1 The different versions of the Global Hunger Index for which the associations of relative changes from 1995 to
2005 were examined include FAO’s proportion of undernourished, the prevalence of wasting and stunting, and
the under-five mortality rate. Regression estimates were not included for wasting and stunting. Correlation
coefficients could be calculated for 66 countries and were highly significant (p-value = 0.000). The three index
dimensions are equally weighted, and stunting and wasting are weighted equally within the child
undernutrition dimension.
2 For the index versions with standardized components, the indicators were standardized using the following
thresholds prior to aggregation: 80% for undernourishment, 65% for underweight, 70% for stunting, 30% for
wasting, and 35% for under-five mortality; the actual maximum values in the data set (Table 5); or the actual
maximum values increased by 10%. The standardized values X (stand.) for each indicator were calculated as
follows: X (stand.) = X/estimated maximum value * 100.
Data sources: See von Grebmer et al. 2014, p.40.
Table 7: Spearman rank correlations for different versions of the Global Hunger Index 1
Global Hunger Index with
non-stand. components including:
underweight
non-stand.
components
including:
stand.
components
including:
stunting
stunting +
wasting
stand. components including:
underweight
stunting
underweight
1.00
stunting
0.95
1.00
stunting + wasting
0.97
0.99
1.00
underweight
0.98
0.93
0.95
1.00
stunting
0.94
0.99
0.98
0.94
1.00
stunting + wasting
0.98
0.96
0.98
0.98
0.97
stunting +
wasting
1.00
1 The different versions of the Global Hunger Index include FAO’s proportion of undernourished and the underfive mortality rate besides the listed child undernutrition indicators. Nationally representative survey data from
the period 2009-2013 retrieved in May 2014 were used for the child undernutrition indicators, regression
estimates were not included. Correlation coefficients could be calculated for 75 countries and were highly
significant (p-value = 0.000). The three index dimensions are equally weighted, and stunting and wasting are
weighted equally within the child undernutrition dimension if used in combination. For the index versions with
standardized components, the indicators were standardized using the following estimated maximum values
prior to aggregation: 80% for undernourishment, 65% for underweight, 70% for stunting, 30% for wasting, and
35% for under-five mortality. The standardized values X (stand.) for each indicator were calculated as follows: X
(stand.) = X/estimated maximum value * 100. Correlation coefficients for standardized and non-standardized
indices with the same component indicators are emboldened.
Data sources: See von Grebmer et al. 2014, p.40.
Table 7 shows that different versions of the GHI with data from 2009-2013 (excluding countries for
which child underweight had to be estimated for this period, and including only countries for which
survey data for stunting and wasting were also available) are highly redundant in terms of their rank
correlations. This is particularly true for indices with the same component indicators with and
18
without standardization. Yet, as discussed above, standardization of index components is common
practice. In the case of the GHI, standardization can balance the contribution of wasting and stunting
to the child undernutrition dimension and decrease the influence of variations in the
undernourishment indicator on trends over time.
19
5 Conclusion
Building on the discussions in previous sections, we come to the conclusion that (1) the child
underweight indicator should be replaced with child stunting and child wasting; (2) the weight of one
third for the child undernutrition dimension should be shared equally between stunting and wasting;
and (3) the component indicators of the index should be standardized prior to aggregation, using
fixed thresholds set above the maximum values observed in the data set.
Stunting and wasting are weighted equally within the child undernutrition dimension to place equal
emphasis on chronic and acute undernutrition.16 Both forms of child undernutrition entail human
suffering, yet, it is not easy to ascertain which condition has the more serious implications. Stunting
at two years was found to be associated with impaired early motor and cognitive development,
shorter adult height, lower attained schooling, reduced adult income and—for women—decreased
birthweight of their offspring (Victora et al. 2008; Black et al. 2013). In a recent prospective cohort
study in Tanzania, McDonald et al. (2013a) found that wasting was associated with deficits in
psychomotor and mental development, independently of stunting. Wasting is a stronger determinant
of under-five mortality than stunting: The hazard ratio was 11.6 for severe wasting and 5.5 for severe
stunting in a pooled analysis of 10 prospective studies in Africa, Asia and South America, and was also
higher for moderate wasting than for moderate stunting (Olofin et al. 2013).17 There is some
indication that repeated episodes of wasting lead to stunting, but the mechanisms are still poorly
understood (Khara and Dolan 2014). In the absence of conclusive evidence about the relative extent
of suffering associated with wasting and stunting, equal weighting of the two indicators is preferred.
Taking into account the thresholds selected for standardization of 80% for undernourishment, 70%
for stunting, 30% for wasting, and 35% for under-five mortality, results in the following formula for a
revised GHI:
GHI = 1/3*(PUN/80*100) + 1/6*(CST/70*100) + 1/6*(CWA/30*100) + 1/3*(CM/35*100)
with
GHI:
PUN:
CST:
CWA:
CM:
Global Hunger Index
proportion of the population that is undernourished (in %)
prevalence of stunting in children younger than five years (in %)
prevalence of wasting in children younger than five years (in %)
proportion of children dying before the age of five years (in %)
Alternatively, the calculations can be shown in two steps:
Standardization of component indicators:
PUN (stand.) = PUN/80*100
CST (stand.) = CST/70*100
CWA (stand.) = CWA/30*100
CM (stand.) = CM/35*100
16
Stunting has higher correlations with the undernourishment indicator and the under-five mortality rate than
wasting; this may be related to the short-term fluctuations of acute undernutrition and the challenges of
obtaining national prevalence estimates that are representative of all seasons, and does not mandate the use
of different weights for the two indicators.
17
Children who were moderately or severely stunted and moderately or severely wasted (and thereby, almost
inevitably, also underweight: only 6 out of 2959 children who were both stunted and wasted in the quoted
study were not underweight) had an even higher hazard ratio of 12.3 (McDonald et al. 2013b). This is another
argument to include both wasting and stunting in the GHI: High prevalence rates of these two forms of
undernutrition in the same population make their co-occurrence in the same individuals more likely, and, if
both indicators are considered, they increase the GHI scores compared to countries where only one form is
highly prevalent.
20
Aggregation of component indicators:
GHI = 1/3*PUN (stand.) + 1/6*CST (stand.) + 1/6*CWA (stand.) + 1/3*CM (stand.)
The new GHI formula was utilized to calculate GHI scores for the first time in 2015 (von Grebmer et
al. 2015). The cutoffs for classifying countries into categories of low, moderate, serious, alarming and
extremely alarming hunger were adapted because the component indicators of the index were
standardized as described in this paper, affecting the range of GHI scores relative to years past (see
Chapter 1 of the 2015 Global Hunger Index report, von Grebmer et al. 2015). Concluding, Figure 4
illustrates the composition and weighting of the revised GHI, lists the main data sources, and
summarizes the reasons for including each dimension and indicator.
21
Figure 4: Overview of dimensions and indicators of the revised Global Hunger Index
3 Dimensions
Inadequate food supply
(FAO)
4 Indicators
1/3 Calorie
deficiency
Reasons for inclusion
•
•
•
•
Global
Hunger
Index
Child undernutrition
(UNICEF/WHO/World
Bank)
1/6 Stunting
1/6 Wasting
•
•
•
•
Child mortality
(IGME)
1/3 Under-five
mortality rate
•
•
Undernourishment (FAO)
measures insufficient calorie
availability, important indicator
of hunger
Refers to entire population,
children and adults
Lead indicator for international
hunger reduction targets
Goes beyond calorie availability,
considers aspects of diet quality
and utilization
Children are particularly
vulnerable to nutritional
deficiencies
Child undernutrition is sensitive
to inequitable intra-household
distribution
Suggested nutrition indicators
for Sustainable Development
Goals (SDGs)
Most serious consequence of
hunger, children most affected
Wasting and stunting capture
mortality risk of undernutrition
only partially
Improves correlations of index
with micronutrient deficiencies
Source: Authors’ own presentation
22
References
Anand, S., and K. Hanson. 1997. Disability-adjusted life years: a critical review. Journal of Health
Economics 16: 685-702.
Andersson, M., V. Karumbunathan, and M. B. Zimmermann. 2012. Global Iodine Status in 2011 and
Trends over the Past Decade. The Journal of Nutrition 142: 744-750.
Black, R. E., C. G. Victora, S. P. Walker, Z. A. Bhutta, P. Christian, M. de Onis, M. Ezzati, S. GranthamMcGregor, J. Katz, R. Martorell, and R. Uauy. 2013. Maternal and Child Undernutrition and
Overweight in Low-Income and Middle-Income Countries. The Lancet 382 (9890): 427-451.
Black, R. E., L. H. Allen, Z. A. Bhutta, L. E. Caulfield, M. de Onis, M. Ezzati, C. Mathers, and J. Rivera, for
the Maternal and Child Undernutrition Study Group. 2008. Maternal and child undernutrition:
Global and regional exposures and consequences. The Lancet 371 (9608): 243–60.
Black, R. E., S. Cousens, H. L. Johnson, J. E. Lawn, I. Rudan, D. G. Bassani, P. Jha, H. Campbell, C.
Fischer Walker, R. Cibulskis, T. Eisele, L. Liu, and C. Mathers, for the Child Health Epidemiology
Reference Group of WHO and UNICEF. 2010. Global, regional, and national causes of child
mortality in 2008: a systematic analysis. The Lancet 375 (9730): 1969-87.
de Benoist, B., E. McLean, I. Egli, and M. Cogswell. 2008. Worldwide Prevalence of Anaemia 1993–
2005: WHO Global Database on Anaemia. Geneva: World Health Organization.
FAO (Food and Agriculture Organization of the United Nations). 2015. Food Security Indicators
(Updated May 27, 2015). Accessed June 1, 2015. http://faostat3.fao.org/download/D/*/E.
FAO, IFAD (International Fund for Agricultural Development) and WFP (World Food Programme).
2013. The State of Food Insecurity in the World 2013. The multiple dimensions of food security.
Rome.
FAO, IFAD, and WFP. 2015. The State of Food Insecurity in the World 2015. Meeting the 2015
International Hunger Targets: Taking Stock of Uneven Progress. Rome.
FAO and IRD (Institut de Recherche pour le Développement). 2014. Defining a Standard Operational
Indicator of Women’s Dietary Diversity: The Women’s Dietary Diversity Follow-up Project.
Contributors: Y. Martin-Prével, P. Allemand, D. Wiesmann, M. Arimond, T. J. Ballard, M. Deitchler,
M. C. Dop, G. Kennedy, W. T. K. Lee, and M. Moursi. Rome and Montpellier.
IFPRI (International Food Policy Research Institute). 2014. Global Nutrition Report 2014: Actions and
Accountability to Accelerate the World’s Progress on Nutrition. Washington, DC.
IFPRI. 2015. Global Nutrition Report 2015: Actions and Accountability to Accelerate the World’s
Progress on Nutrition. Washington, DC.
IGME (Inter-agency Group for Child Mortality Estimation). 2014. Child Mortality Estimates. Accessed
May 21, 2015. www.childmortality.org
India, Ministry of Women and Child Development, and UNICEF, India. 2014. Rapid Survey on Children
(2013-14).
Khara, T., and C. Dolan. 2014. Technical Briefing Paper: Associations between wasting and stunting,
policy, programming and research implications. Emergency Nutrition Network (ENN), June 2014.
Kothari, M., and N. Abderrahim. 2010. Nutrition Update 2010. Calverton, Maryland, US: ICF Macro.
Liu, L., H. L. Johnson, S. Cousens, J. Perin, S. Scott, J. E. Lawn, I. Rudan, H. Campbell, R. Cibulskis, M. Li,
C. Mathers, and R. E. Black, for the Child Health Epidemiology Reference Group of WHO and
UNICEF. 2012. Global, regional, and national causes of child mortality: an updated systematic
analysis for 2010 with time trends since 2000. The Lancet 379 (9832): 2151–61.
Liu, L., S. Oza, D. Hogan, J. Perin, I. Rudan, J. E. Lawn, S. Cousens, C. Mathers, and R. E. Black. 2015.
Global, regional, and national causes of child mortality in 2000-13, with projections to inform
post-2015 priorities: an updated systematic analysis. The Lancet 385 (9966): 430-40.
23
Martorell, R. 2008. Malnutrition and Hunger, Copenhagen Consensus 2008 Perspective Paper.
http://www.copenhagenconsensus.com/Default.aspx?ID=1322
Masset, E. 2011. A review of hunger indices and methods to monitor country commitment to fighting
hunger. Food Policy 36: S102-S108.
McDonald, C. M., K. P. Manji, R. Kupka, D. C. Bellinger, D. Spiegelman, R. Kisenge, G. Msamanga, W.
W. Fawzi, and C. P. Duggan. 2013a. Stunting and wasting are associated with poorer psychomotor
and mental development in HIV-exposed Tanzanian infants. Journal of Nutrition 143(2): 204-214.
McDonald, C. M., I. Olofin, S. Flaxman, W. W. Fawzi, D. Spiegelman, L. E. Caulfield, R. E. Black, M.
Ezzati, and G. Danaei, for the Nutrition Impact Model Study. 2013b. The effect of multiple
anthropometric deficits on child mortality: meta-analysis of individual data in 10 prospective
studies from developing countries. American Journal of Clinical Nutrition 97(4): 896-901.
MEASURE DHS. 2014. Demographic and Health Surveys. Calverton, Maryland, USA. Accessed May 1,
2014. www.measuredhs.com
Menon, P., A. Deolalikar, and A. Bhaskar. 2009. India State Hunger Index: Comparisons of Hunger
Across States. International Food Policy Research Institute, Welthungerhilfe, and University of
California, Riverside. Washington, DC, Bonn, and Riverside.
Muthayya, S., J. Hyun Rah, J. D. Sugimoto, F. F. Roos, K. Kraemer, and R. E. Black. 2013. The Global
Hidden Hunger Indices and Maps: An Advocacy Tool for Action. PLoS ONE 8(6): e67860.
Newby, H. 2014. Data Gaps for Low Birth Weight. Technical Note 9. In: IFPRI. 2014. Global Nutrition
Report 2014: Actions and Accountability to Accelerate the World’s Progress on Nutrition.
Washington, DC.
Noorbakhsh, F. 1998. The Human Development Index: Some Technical Issues and Alternative Indices.
Journal of International Development 10: 589-605.
Olofin, I., C. M. McDonald, M. Ezzati, S. Flaxman, R. E. Black, W. W. Fawzi, L. E. Caulfield, and G.
Danaei for the Nutrition Impact Model Study (anthropometry cohort pooling). 2013. Associations
of suboptimal growth with all-cause and cause-specific mortality in children under five years: a
pooled analysis of ten prospective studies. PLoS One 8(5): e64636.
Ryten, J. 2000. Should there be a Human Development Index? Paper presented at the IAOS 2000
conference on "Statistics, Human Rights and Development", Montreux, September 4-8, 2000.
Schmidt, E., and P. Dorosh. 2009. A Sub-National Hunger Index for Ethiopia: Assessing Progress in
Region-Level Outcomes. Ethiopia Strategy Support Program 2 (ESSP2) Discussion Paper No. 005.
International Food Policy Research Institute. Addis Ababa and Washington, DC.
Smith, L. C., and L. Haddad. 2015. Reducing Child Undernutrition: Past Drivers and Priorities for the
Post-MDG Era. World Development 68 (April 2015): 180–204.
Stevens, G. A., M. M. Finucane, L. M. De-Regil, C. J. Paciorek, S. R. Flaxman, F. Branca, J. P. PeñaRosas, et al. 2013. Global, Regional, and National Trends in Haemoglobin Concentration and
Prevalence of Total and Severe Anaemia in Children and Pregnant and Non-Pregnant Women for
1995-2011: A Systematic Analysis of Population-Representative Data. The Lancet Global Health 1
(1): e16-25. http://dx.doi.org/10.1016/s2214-109x(13)70001-9
Swift, J. 1993. Understanding and Preventing Famine and Famine Mortality. IDS (Institute for
Development Studies) Bulletin 24.4. Brighton.
Szilágyi, G. 2000. International comparison theory - human development. Paper presented at the
IAOS 2000 conference on "Statistics, Human Rights and Development", Montreux, September 4-8,
2000.
Tomkins, A., and F. Watson. 1989. Malnutrition and Infection - a Review. UN ACC/SCN (United
Nations Administrative Committee on Coordination/Subcommittee on Nutrition) Nutrition Policy
Discussion Paper No.5. Geneva.
UNDP (United Nations Development Programme). 2009. Human Development Report. Overcoming
barriers: Human mobility and development. New York.
24
UNDP. 2010. Human Development Report. The Real Wealth of Nations: Pathways to Human
Development. New York.
UNICEF (United Nations Children’s Fund). 1990. Strategy for Improved Nutrition of Children and
Women in Developing Countries. New York.
UNICEF. 2014. Childinfo: Multiple Indicator Cluster Surveys (MICS). Accessed May 30, 2014.
www.childinfo.org/mics_available.html.
UNICEF/WHO (World Health Organization)/World Bank. 2013. UNICEF-WHO-The World Bank: 2012
Joint Child Malnutrition Estimates-Levels and Trends. New York, Geneva, and Washington, DC.
www.who.int/nutgrowthdb/estimates2012
United Nations. 2008. Official List of MDG Indicators.
http://unstats.un.org/unsd/mdg/Host.aspx?Content=Indicators/OfficialList.htm
UNSCN (United Nations Standing Committee on Nutrition). 2014. Priority Nutrition Indicators for the
Post-2015 Sustainable Development Goals. Policy Brief.
http://www.unscn.org/files/Publications/Policy_brief_Priority_Nutrition_Indicators_for_the_Post
-2015_SDGs.pdf
Victora, C. G., L. Adair, C. Fall, P. C. Hallal, R. Martorell, L. Richter, and H. Singh Sachdev, for the
Maternal and Child Undernutrition Study Group. 2008. Maternal and child undernutrition:
Consequences for adult health and human capital. The Lancet 371 (9609): 340–57.
von Grebmer, K., A. Saltzman, E. Birol, D. Wiesmann, N. Prasai, S. Yin, Y. Yohannes, P. Menon, J.
Thompson, and A. Sonntag. 2014. 2014 Global Hunger Index: The Challenge of Hidden Hunger.
Welthungerhilfe, International Food Policy Research Institute, and Concern Worldwide. Bonn,
Washington, DC, and Dublin.
von Grebmer, K., J. Bernstein, A. de Waal, N. Prasai, S. Yin, and Y. Yohannes. 2015. 2015 Global
Hunger Index: Armed Conflict and the Challenge of Hunger. Welthungerhilfe, International Food
Policy Research Institute, and Concern Worldwide. Bonn, Washington, DC, and Dublin.
WFP Nepal (World Food Programme Nepal). 2009. A sub-regional hunger index for Nepal. Nepal
Food Security Monitoring System. Kathmandu, Nepal.
WHO (World Health Organization). 1995. Physical status: the use and interpretation of
anthropometry. Report of a WHO Expert Committee. Technical Report Series No. 854. Geneva.
WHO. 2009a. Global Prevalence of Vitamin A Deficiency in Populations at Risk 1995-2005: WHO
Global Database on Vitamin A deficiency. www.who.int/vmnis/vitamina/en/.
WHO. 2009b. Death and DALY estimates for 2004 by cause for WHO Member States. Accessed
September 5, 2014.
http://www.who.int/healthinfo/global_burden_disease/estimates_country/en/
WHO. 2010a. Indicators for Assessing Infant and Young Child Feeding Practices. Part 1: Definitions.
Geneva.
WHO. 2010b. Indicators for Assessing Infant and Young Child Feeding Practices. Part 3: Country
profiles. Geneva.
WHO. 2012. Document WHA65/2012/REC1 Annex 2 “Comprehensive Implementation Plan on
Maternal, Infant and Young Child Nutrition”. Sixty-Fifth World Health Assembly, Geneva, May 2126 2012. http://www.who.int/nutrition/topics/WHA65.6_annex2_en.pdf
WHO. 2014a. The WHO Global Database on Child Growth and Malnutrition. Accessed May 26, 2014.
www.who.int/nutgrowthdb/en/
WHO. 2014b. Tracking progress in achieving the global nutrition targets - May 2014.
http://www.who.int/nutrition/targets_indicators_may2014v2.pdf
Wiesmann, D. 2004. An International Nutrition Index: Concept and Analyses of Food Insecurity and
Undernutrition at Country Levels. Frankfurt a. M.: Peter Lang.
25
Wiesmann, D. 2006. A Global Hunger Index: Measurement concept, ranking of countries and trends.
Food Consumption and Nutrition Division Discussion Paper 212. Washington, DC: International
Food Policy Research Institute.
Wiesmann, D., J. von Braun, and T. Feldbrügge. 2000. An International Nutrition Index: Successes and
Failures in Addressing Hunger and Malnutrition. ZEF Discussion Papers on Development Policy No.
26. Zentrum für Entwicklungsforschung (ZEF) [Center for Development Research]. Bonn.
Wiesmann, D., L. Weingärtner, and I. Schöninger. 2006. The Challenge of Hunger: Global Hunger
Index: Facts, determinants, and trends. Deutsche Welthungerhilfe and International Food Policy
Research Institute. Bonn and Washington, DC.
World Bank. 2014. World Bank Open Data. http://data.worldbank.org/
26
Appendix A: Nutrition indicators proposed for the
Sustainable Development Goals (SDGs) and not
recommended for inclusion in the Global Hunger Index
Eight nutrition indicators have been proposed for consideration in the SDGs: the six indicators
endorsed by the World Health Assembly, and two additional indicators—one on dietary diversity,
and the other one on policy (Table 1). Only two indicators, namely stunting and wasting in children
under five, are included in the revised GHI formula.
Two of the eight indicators were not considered for the GHI for conceptual reasons:
The percentage of national budget allocated to nutrition should not be included in the GHI because
it is an input to and not an outcome of good nutrition. It is recommended to retain the focus of the
GHI on outcomes closely related to human well-being (dietary energy deficiency in the population,
child undernutrition, and child mortality).
The prevalence of overweight in children under five is an important indicator for monitoring
nutrition. Yet, it should not become part of the GHI because the index is about capturing deficiencies
and deprivation and not excess. If an indicator of overweight or obesity were integrated into the GHI,
it could no longer be truthfully called a “hunger index”. The correlation coefficients in Table 3 show
that child overweight is either not at all or negatively associated with various indicators of
micronutrient deficiencies.18 Indicators of overweight or obesity might better be used to construct a
nutrition or overnutrition index.
Four other of the proposed eight indicators have serious data gaps and for some, comparability
across countries is also limited:
The percentage of women of reproductive age with anemia is a proxy indicator and not a precise
measure of iron deficiency because there are many other causes of anemia besides iron deficiency,
including other nutritional deficits. The Global Nutrition Report (GNR) assesses that only 5 out of 185
countries are currently on track to meet the WHA target to halve anemia in women of reproductive
age by 2025 (IFPRI 2014). Monitoring this indicator and stimulating political will to reduce anemia
among women is therefore a high priority.
The percentage of women of reproductive age with anemia would be a great candidate for inclusion
into the GHI. The index could be extended to include a fourth dimension for micronutrient
deficiencies or hidden hunger in addition to the existing three dimensions, or anemia in women of
reproductive age could be integrated into one nutrition dimension together with indicators of child
undernutrition.
However, limited data availability is a serious obstacle: Among the 120 countries for which the 2014
GHI could be computed, only 29 have data on anemia in women of reproductive age from nationally
representative surveys for 2010 or later, and only 38 countries have such data for 2008 or later
(Stevens et al. 2013). If the percentage of women of reproductive age with anemia was included into
the GHI, values would have to be estimated by means of regression models for about 90 countries
for the 2015 GHI (reference period 2010-2016). The estimation would be especially challenging for
the 40 out of 120 countries that never had any survey on anemia in women of reproductive age and
18
The observed significant negative correlations indicate that night blindness among pregnant women and
anemia among preschool children are less prevalent in countries with a higher proportion of overweight
children under five. This does not imply causality, but suggests that rising child overweight does not necessarily
mean a greater extent of hidden hunger.
27
would most likely result in shaky estimates for these countries. This would make the ranking based
on the 2015 GHI less defensible.19
It is recommended to reconsider including an indicator of anemia among women of reproductive age
into the GHI after more survey data have become available. The fact that a WHA target has been
defined for this indicator will hopefully spur greater efforts to collect data.
The percentage of women, 15-49 years of age, who consume at least 5 out of 10 defined food
groups, is based on a 10-food group dietary diversity indicator that proxies micronutrient adequacy
among women of reproductive age. The new global indicator captures aspects of diet quality and
was endorsed by a wide range of stakeholders at a consensus meeting convened by the Food and
Nutrition Technical Assistance Project (FANTA) and FAO in July 2014.20
So far, data from nationally representative surveys for this indicator are virtually absent. Also, the
correlation between the percentage of women consuming at least 5 out of 10 defined food groups
and the percentage of women above a 50%, 60% or 70% cutoff for micronutrient adequacy was not
very strong for the limited set of cross-sectional data from diverse settings and countries that is
available (FAO and IRD 2014). Much greater data availability would be required and the validity of
the indicator for cross-country comparisons of diet quality should be further investigated before
considering it for the GHI.
The percentage of infants born with low birth weight is related to fetal undernutrition and therefore
provides information about an important time period within the nutritionally critical 1,000 days after
conception, and also about mothers’ nutritional status during pregnancy. In principle, it would be
worthwhile to examine statistically if this indicator could meaningfully complement other indicators
of child undernutrition that have been selected for the GHI.
However, data on low birth weight are sparse and inconsistent. Globally, only about half of all babies
are weighed at birth, and the share of weighed babies varies greatly across countries. Estimates of
the prevalence of low birth weight are not always comparable across time and across countries, and
improved adjustment methods need to be developed (Newby 2014).
At present, both the lack of data and the limited comparability across countries prevent the inclusion
of the percentage of infants born with low birth weight into the GHI. If more data become available
and reliable adjustment methods can be developed, the potential benefits of including this indicator
should be examined.
The percentage of infants less than 6 months of age who are exclusively breastfed is an indicator of
child feeding. It can be considered an input to child nutrition rather than an outcome measure.
Including this indicator into the GHI would be possible, yet, taking into account the type of outcome
indicators selected for the GHI, this does not have high priority.
The data collection method for exclusive breastfeeding among infants under 6 months limits the
comparability of the indicator across countries. The indicator is calculated as the proportion of
infants 0-5 months of age who received only breast milk during the previous day among all infants 05 months of age. Because some infants are given other liquids than breast milk irregularly, using the
previous day as recall period will lead to an overestimation of exclusively breastfed infants (WHO
2010a). The degree of overestimation will vary across countries because it is more common in some
settings to give other liquids than breast milk irregularly than in others.
19
In a section on data gaps, the GNR states that “More survey-based micronutrient data are
required. Anemia data are based on models, and modeled data at the country level may not be
considered meaningful or credible” (IFPRI 2014, p.21).
20
http://www.fantaproject.org/sites/default/files/resources/Introduce-MDD-W-indicator-brief-Sep2014.pdf
28
Data availability is also still limited: World Bank (2014) presented data from 2010 or later for only 51
out of the 120 countries for which the 2014 GHI could be computed, which means that data for
almost 70 countries would have to be estimated.
29
Appendix B: Reasons to include child mortality in the Global
Hunger Index
Three main arguments to include the child mortality dimension in the GHI can be brought forward:
1. Premature death is the most serious consequence of hunger, and children are more vulnerable
than adults.
Children are particularly vulnerable to nutritional deficiencies because of their fragile immune
systems: They are at a higher risk than adults to get caught in a vicious cycle of inadequate food
intake, infection, loss of appetite and poor nutrient absorption, often with fatal consequences
(UNICEF 1990; Tomkins and Watson 1989). It is true that food insecurity and undernutrition also tend
to raise adult mortality rates, especially during famines (Swift 1993). Yet, when comparing across
countries, mortality among adults varies to a larger extent according to factors that are not linked to
undernutrition, such as other lifestyle aspects, hazardous occupations, or active participation in wars.
Chronic diseases that are more strongly associated with unbalanced food intake than with
undernutrition, such as cardiovascular diseases and cancer, also play a larger role for adult mortality
than for child mortality (WHO 2009b). Therefore, child mortality is preferred as a proxy for
premature mortality associated with undernutrition.
2. The contribution of nutritional disorders to child mortality is only partially captured by stunting
and wasting.
Underweight also contributes to child mortality, and so do zinc deficiency, vitamin A deficiency, fetal
growth restriction and suboptimum breastfeeding (Black et al. 2013). At present, indicators of
micronutrient deficiencies and fetal growth restriction (that is, low birth weight) cannot be included
into the GHI because of limited data availability and/or comparability. It is very likely that the
contribution of micronutrient deficiencies to child mortality is underestimated, because data and
sometimes also indicators are lacking for most micronutrient deficiencies.
There is an overlap in the contribution of the various nutritional factors to child mortality: If the
percentages are added up (Table B-1, numbers based on UN prevalences, second data column,), the
result amounts to 76.5%, although all nutritional disorders cause only an estimated 44.7% of all
deaths of children under five. Children can be affected by more than one nutritional disorder, and
this has to be taken into account. Because fetal growth restriction and suboptimum breastfeeding in
neonates (that is, infants <1 month) does not overlap with stunting, wasting, and underweight
among children 1-59 months in terms of age group, its joint contribution to 19.4% of all child deaths
is independent of the indicated contribution of stunting, wasting, and underweight. This means that
fetal growth restriction and suboptimum breastfeeding in neonates jointly cause 43% of all child
mortality that is attributable to nutritional disorders (19.4% divided by 44.7%).
3. Child mortality has higher correlations overall with micronutrient deficiencies than the other
components of the GHI.
Among the indicators included in the GHI up to 2014, child mortality had the highest correlations
overall with indicators of micronutrient deficiencies (Figure B-1). This is particularly true for anemia
in preschool children and pregnant women and low serum retinol in preschool children. For night
blindness among pregnant women, underweight had slightly higher correlations than child mortality,
but the difference was not very large.21 Including child mortality in the GHI therefore improves the
correlations of the index with indicators of micronutrient deficiencies. This is a desirable outcome; it
would be even better to include indicators of micronutrient deficiencies in the index, yet, this is not
possible to date due to the scarcity of data. As long as the availability of data on micronutrient
21
Low dietary diversity among young children is no measure of micronutrient deficiencies, but only a proxy
indicator of diet quality.
30
deficiencies does not increase significantly, the under-five mortality rate may serve as a kind of proxy
indicator for these deficiencies.
Table B-1: Global deaths in children younger than five years attributable to nutritional
disorders
Source: Black et al. 2013
31
Figure B-1: Correlations of the Global Hunger Index with measures of hidden hunger
Spearman rank correlation coefficients can range from 0 (no association) to 1 (perfect association). All correlations with the GHI are statistically significant at p-value < 0.01. For
the GHI components, solid color indicates significance at p < 0.05. Nationally representative survey data were used for indicators of micronutrient deficiencies and diet
diversity. The latest available data were matched with the GHI and its components using the year of the survey and the GHI reference periods. N indicates the number of
countries for which the correlation coefficients could be computed.
Definitions and data sources: Low dietary diversity: Proportion of children 6–23 months who consume fewer than four out of seven food groups (grains, roots and tubers;
legumes and nuts; dairy products; flesh foods; eggs; vitamin-A rich fruits and vegetables; other fruits and vegetables) (WHO 2010b; Kothari and Abderrahim 2010). Anemia:
Proportion of preschool-age children whose hemoglobin level is less than 110 grams per liter, and proportion of pregnant women whose hemoglobin level is less than 110
grams per liter (World Bank 2014; MEASURE DHS 2014; de Benoist et al. 2008). Vitamin A deficiency: Proportion of preschool-age children with night blindness, proportion of
pregnant women with night blindness, and proportion of preschool-age children whose serum retinol level is less than 0.70 micromole per liter (WHO 2009a).
Source: von Grebmer et al. 2014
32
ZEF Working Paper Series, ISSN 1864-6638
Department of Political and Cultural Change
Center for Development Research, University of Bonn
Editors: Christian Borgemeister, Joachim von Braun, Manfred Denich, Solvay Gerke, Eva Youkhana and Till
Stellmacher
1.
2.
3.
4.
5.
6.
7.
8.
8.a
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Evers, Hans-Dieter and Solvay Gerke (2005). Closing the Digital Divide: Southeast Asia’s Path Towards a
Knowledge Society.
Bhuiyan, Shajahan and Hans-Dieter Evers (2005). Social Capital and Sustainable Development: Theories and
Concepts.
Schetter, Conrad (2005). Ethnicity and the Political Reconstruction of Afghanistan.
Kassahun, Samson (2005). Social Capital and Community Efficacy. In Poor Localities of Addis Ababa Ethiopia.
Fuest, Veronika (2005). Policies, Practices and Outcomes of Demand-oriented Community Water
Supply in Ghana: The National Community Water and Sanitation Programme 1994 – 2004.
Menkhoff, Thomas and Hans-Dieter Evers (2005). Strategic Groups in a Knowledge Society: Knowledge
Elites as Drivers of Biotechnology Development in Singapore.
Mollinga, Peter P. (2005). The Water Resources Policy Process in India: Centralisation, Polarisation and New
Demands on Governance.
Evers, Hans-Dieter (2005). Wissen ist Macht: Experten als Strategische Gruppe.
Evers, Hans-Dieter and Solvay Gerke (2005). Knowledge is Power: Experts as Strategic Group.
Fuest, Veronika (2005). Partnerschaft, Patronage oder Paternalismus? Eine empirische Analyse
der Praxis universitärer Forschungskooperation mit Entwicklungsländern.
Laube, Wolfram (2005). Promise and Perils of Water Reform: Perspectives from Northern Ghana.
Mollinga, Peter P. (2004). Sleeping with the Enemy: Dichotomies and Polarisation in Indian Policy Debates on
the Environmental and Social Effects of Irrigation.
Wall, Caleb (2006). Knowledge for Development: Local and External Knowledge in Development Research.
Laube, Wolfram and Eva Youkhana (2006). Cultural, Socio-Economic and Political Con-straints for Virtual
Water Trade: Perspectives from the Volta Basin, West Africa.
Hornidge, Anna-Katharina (2006). Singapore: The Knowledge-Hub in the Straits of Malacca.
Evers, Hans-Dieter and Caleb Wall (2006). Knowledge Loss: Managing Local Knowledge in Rural Uzbekistan.
Youkhana, Eva; Lautze, J. and B. Barry (2006). Changing Interfaces in Volta Basin Water Management:
Customary, National and Transboundary.
Evers, Hans-Dieter and Solvay Gerke (2006). The Strategic Importance of the Straits of Malacca for
World Trade and Regional Development.
Hornidge, Anna-Katharina (2006). Defining Knowledge in Germany and Singapore: Do the Country-Specific
Definitions of Knowledge Converge?
Mollinga, Peter M. (2007). Water Policy – Water Politics: Social Engineering and Strategic Action in Water
Sector Reform.
Evers, Hans-Dieter and Anna-Katharina Hornidge (2007). Knowledge Hubs Along the Straits of Malacca.
Sultana, Nayeem (2007). Trans-National Identities, Modes of Networking and Integration in a MultiCultural Society. A Study of Migrant Bangladeshis in Peninsular Malaysia.
Yalcin, Resul and Peter M. Mollinga (2007). Institutional Transformation in Uzbekistan’s Agricultural
and Water Resources Administration: The Creation of a New Bureaucracy.
Menkhoff, T.; Loh, P. H. M.; Chua, S. B.; Evers, H.-D. and Chay Yue Wah (2007). Riau Vegetables for
Singapore Consumers: A Collaborative Knowledge-Transfer Project Across the Straits of Malacca.
Evers, Hans-Dieter and Solvay Gerke (2007). Social and Cultural Dimensions of Market Expansion.
25. Obeng, G. Y.; Evers, H.-D.; Akuffo, F. O., Braimah, I. and A. Brew-Hammond (2007). Solar PV Rural
Electrification and Energy-Poverty Assessment in Ghana: A Principal Component Analysis.
26. Eguavoen, Irit; E. Youkhana (2008). Small Towns Face Big Challenge. The Management of Piped
Systems after the Water Sector Reform in Ghana.
27. Evers, Hans-Dieter (2008). Knowledge Hubs and Knowledge Clusters: Designing a Knowledge Architecture for
Development
28. Ampomah, Ben Y.; Adjei, B. and E. Youkhana (2008). The Transboundary Water Resources Management
Regime of the Volta Basin.
29. Saravanan.V.S.; McDonald, Geoffrey T. and Peter P. Mollinga (2008). Critical Review of Integrated
Water Resources Management: Moving Beyond Polarised Discourse.
30. Laube, Wolfram; Awo, Martha and Benjamin Schraven (2008). Erratic Rains and Erratic Markets:
Environmental change, economic globalisation and the expansion of shallow groundwater irrigation in West
Africa.
31. Mollinga, Peter P. (2008). For a Political Sociology of Water Resources Management.
32. Hauck, Jennifer; Youkhana, Eva (2008). Histories of water and fisheries management in Northern Ghana.
33. Mollinga, Peter P. (2008). The Rational Organisation of Dissent. Boundary concepts, boundary objects and
boundary settings in the interdisciplinary study of natural resources management.
34. Evers, Hans-Dieter; Gerke, Solvay (2009). Strategic Group Analysis.
35. Evers, Hans-Dieter; Benedikter, Simon (2009). Strategic Group Formation in the Mekong Delta - The
Development of a Modern Hydraulic Society.
36. Obeng, George Yaw; Evers, Hans-Dieter (2009). Solar PV Rural Electrification and Energy-Poverty: A
Review and Conceptual Framework With Reference to Ghana.
37. Scholtes, Fabian (2009). Analysing and explaining power in a capability perspective.
38. Eguavoen, Irit (2009). The Acquisition of Water Storage Facilities in the Abay River Basin, Ethiopia.
39. Hornidge, Anna-Katharina; Mehmood Ul Hassan; Mollinga, Peter P. (2009). ‘Follow the Innovation’ – A
joint experimentation and learning approach to transdisciplinary innovation research.
40. Scholtes, Fabian (2009). How does moral knowledge matter in development practice, and how can it be
researched?
41. Laube, Wolfram (2009). Creative Bureaucracy: Balancing power in irrigation administration in northern
Ghana.
42. Laube, Wolfram (2009). Changing the Course of History? Implementing water reforms in Ghana and South
Africa.
43. Scholtes, Fabian (2009). Status quo and prospects of smallholders in the Brazilian sugarcane and
ethanol sector: Lessons for development and poverty reduction.
44. Evers, Hans-Dieter; Genschick, Sven; Schraven, Benjamin (2009). Constructing Epistemic Landscapes:
Methods of GIS-Based Mapping.
45. Saravanan V.S. (2009). Integration of Policies in Framing Water Management Problem: Analysing Policy
Processes using a Bayesian Network.
46. Saravanan V.S. (2009). Dancing to the Tune of Democracy: Agents Negotiating Power to Decentralise Water
Management.
47. Huu, Pham Cong; Rhlers, Eckart; Saravanan, V. Subramanian (2009). Dyke System Planing: Theory and
Practice in Can Tho City, Vietnam.
48. Evers, Hans-Dieter; Bauer, Tatjana (2009). Emerging Epistemic Landscapes: Knowledge Clusters in Ho Chi
Minh City and the Mekong Delta.
49. Reis, Nadine; Mollinga, Peter P. (2009). Microcredit for Rural Water Supply and Sanitation in the
Mekong Delta. Policy implementation between the needs for clean water and ‘beautiful latrines’.
50. Gerke, Solvay; Ehlert, Judith (2009). Local Knowledge as Strategic Resource: Fishery in the Seasonal
Floodplains of the Mekong Delta, Vietnam
51. Schraven, Benjamin; Eguavoen, Irit; Manske, Günther (2009). Doctoral degrees for capacity
development: Results from a survey among African BiGS-DR alumni.
52. Nguyen, Loan (2010). Legal Framework of the Water Sector in Vietnam.
53. Nguyen, Loan (2010). Problems of Law Enforcement in Vietnam. The Case of Wastewater Management in
Can Tho City.
54. Oberkircher, Lisa et al. (2010). Rethinking Water Management in Khorezm, Uzbekistan. Concepts and
Recommendations.
55. Waibel, Gabi (2010). State Management in Transition: Understanding Water Resources Management in
Vietnam.
56. Saravanan V.S.; Mollinga, Peter P. (2010). Water Pollution and Human Health. Transdisciplinary Research
on Risk Governance in a Complex Society.
57. Vormoor, Klaus (2010). Water Engineering, Agricultural Development and Socio-Economic Trends in the
Mekong Delta, Vietnam.
58. Hornidge, Anna-Katharina; Kurfürst, Sandra (2010). Envisioning the Future, Conceptualising Public Space.
Hanoi and Singapore Negotiating Spaces for Negotiation.
59. Mollinga, Peter P. (2010). Transdisciplinary Method for Water Pollution and Human Health Research.
60. Youkhana, Eva (2010). Gender and the development of handicraft production in rural Yucatán/Mexico.
61. Naz, Farhat; Saravanan V. Subramanian (2010). Water Management across Space and Time in India.
62. Evers, Hans-Dieter; Nordin, Ramli, Nienkemoer, Pamela (2010). Knowledge Cluster Formation in Peninsular
Malaysia: The Emergence of an Epistemic Landscape.
63. Mehmood Ul Hassan; Hornidge, Anna-Katharina (2010). ‘Follow the Innovation’ – The second year of a joint
experimentation and learning approach to transdisciplinary research in Uzbekistan.
64. Mollinga, Peter P. (2010). Boundary concepts for interdisciplinary analysis of irrigation water management in
South Asia.
65. Noelle-Karimi, Christine (2006). Village Institutions in the Perception of National and International Actors in
Afghanistan. (Amu Darya Project Working Paper No. 1)
66. Kuzmits, Bernd (2006). Cross-bordering Water Management in Central Asia. (Amu Darya Project Working
Paper No. 2)
67. Schetter, Conrad; Glassner, Rainer; Karokhail, Masood (2006). Understanding Local Violence. Security
Arrangements in Kandahar, Kunduz and Paktia. (Amu Darya Project Working Paper No. 3)
68. Shah, Usman (2007). Livelihoods in the Asqalan and Sufi-Qarayateem Canal Irrigation Systems in the Kunduz
River Basin. (Amu Darya Project Working Paper No. 4)
69. ter Steege, Bernie (2007). Infrastructure and Water Distribution in the Asqalan and Sufi-Qarayateem Canal
Irrigation Systems in the Kunduz River Basin. (Amu Darya Project Working Paper No. 5)
70. Mielke, Katja (2007). On The Concept of ‘Village’ in Northeastern Afghanistan. Explorations from Kunduz
Province. (Amu Darya Project Working Paper No. 6)
71. Mielke, Katja; Glassner, Rainer; Schetter, Conrad; Yarash, Nasratullah (2007). Local Governance in Warsaj and
Farkhar Districts. (Amu Darya Project Working Paper No. 7)
72. Meininghaus, Esther (2007). Legal Pluralism in Afghanistan. (Amu Darya Project Working Paper No. 8)
73. Yarash, Nasratullah; Smith, Paul; Mielke, Katja (2010). The fuel economy of mountain villages in
Ishkamish and Burka (Northeast Afghanistan). Rural subsistence and urban marketing patterns. (Amu
Darya Project Working Paper No. 9)
74. Oberkircher, Lisa (2011). ‘Stay – We Will Serve You Plov!’. Puzzles and pitfalls of water research in rural
Uzbekistan.
75. Shtaltovna, Anastasiya; Hornidge, Anna-Katharina; Mollinga, Peter P. (2011). The Reinvention of Agricultural
Service Organisations in Uzbekistan – a Machine-Tractor Park in the Khorezm Region.
76. Stellmacher, Till; Grote, Ulrike (2011). Forest Coffee Certification in Ethiopia: Economic Boon or Ecological
Bane?
77. Gatzweiler, Franz W.; Baumüller, Heike; Ladenburger, Christine; von Braun, Joachim (2011). Marginality.
Addressing the roots causes of extreme poverty.
78. Mielke, Katja; Schetter, Conrad; Wilde, Andreas (2011). Dimensions of Social Order: Empirical Fact, Analytical
Framework and Boundary Concept.
79. Yarash, Nasratullah; Mielke, Katja (2011). The Social Order of the Bazaar: Socio-economic embedding of
Retail and Trade in Kunduz and Imam Sahib
80. Baumüller, Heike; Ladenburger, Christine; von Braun, Joachim (2011). Innovative business approaches for the
reduction of extreme poverty and marginality?
81. Ziai, Aram (2011). Some reflections on the concept of ‘development’.
82. Saravanan V.S., Mollinga, Peter P. (2011). The Environment and Human Health - An Agenda for Research.
83. Eguavoen, Irit; Tesfai, Weyni (2011). Rebuilding livelihoods after dam-induced relocation in Koga, Blue Nile
basin, Ethiopia.
84. Eguavoen, I., Sisay Demeku Derib et al. (2011). Digging, damming or diverting? Small-scale irrigation in the
Blue Nile basin, Ethiopia.
85. Genschick, Sven (2011). Pangasius at risk - Governance in farming and processing, and the role of different
capital.
86. Quy-Hanh Nguyen, Hans-Dieter Evers (2011). Farmers as knowledge brokers: Analysing three cases from
Vietnam’s Mekong Delta.
87. Poos, Wolf Henrik (2011). The local governance of social security in rural Surkhondarya, Uzbekistan. PostSoviet community, state and social order.
88. Graw, Valerie; Ladenburger, Christine (2012). Mapping Marginality Hotspots. Geographical Targeting for
Poverty Reduction.
89. Gerke, Solvay; Evers, Hans-Dieter (2012). Looking East, looking West: Penang as a Knowledge Hub.
90. Turaeva, Rano (2012). Innovation policies in Uzbekistan: Path taken by ZEFa project on innovations in the
sphere of agriculture.
91. Gleisberg-Gerber, Katrin (2012). Livelihoods and land management in the Ioba Province in south-western
Burkina Faso.
92. Hiemenz, Ulrich (2012). The Politics of the Fight Against Food Price Volatility – Where do we stand and where
are we heading?
93. Baumüller, Heike (2012). Facilitating agricultural technology adoption among the poor: The role of service
delivery through mobile phones.
94. Akpabio, Emmanuel M.; Saravanan V.S. (2012). Water Supply and Sanitation Practices in Nigeria:
Applying
Local Ecological Knowledge to Understand Complexity.
95. Evers, Hans-Dieter; Nordin, Ramli (2012). The Symbolic Universe of Cyberjaya, Malaysia.
96. Akpabio, Emmanuel M. (2012). Water Supply and Sanitation Services Sector in Nigeria: The Policy Trend and
Practice Constraints.
97. Boboyorov, Hafiz (2012). Masters and Networks of Knowledge Production and Transfer in the Cotton Sector
of Southern Tajikistan.
98. Van Assche, Kristof; Hornidge, Anna-Katharina (2012). Knowledge in rural transitions - formal and informal
underpinnings of land governance in Khorezm.
99. Eguavoen, Irit (2012). Blessing and destruction. Climate change and trajectories of blame in Northern Ghana.
100. Callo-Concha, Daniel; Gaiser, Thomas and Ewert, Frank (2012). Farming and cropping systems in the West
African Sudanian Savanna. WASCAL research area: Northern Ghana, Southwest Burkina Faso and Northern
Benin.
101. Sow, Papa (2012). Uncertainties and conflicting environmental adaptation strategies in the region of the Pink
Lake, Senegal.
102. Tan, Siwei (2012). Reconsidering the Vietnamese development vision of “industrialisation and modernisation
by 2020”.
103. Ziai, Aram (2012). Postcolonial perspectives on ‘development’.
104. Kelboro, Girma; Stellmacher, Till (2012). Contesting the National Park theorem? Governance and land use in
Nech Sar National Park, Ethiopia.
105. Kotsila, Panagiota (2012). “Health is gold”: Institutional structures and the realities of health access in the
Mekong Delta, Vietnam.
106. Mandler, Andreas (2013). Knowledge and Governance Arrangements in Agricultural Production: Negotiating
Access to Arable Land in Zarafshan Valley, Tajikistan.
107. Tsegai, Daniel; McBain, Florence; Tischbein, Bernhard (2013). Water, sanitation and hygiene: the missing link
with agriculture.
108. Pangaribowo, Evita Hanie; Gerber, Nicolas; Torero, Maximo (2013). Food and Nutrition Security Indicators: A
Review.
109. von Braun, Joachim; Gerber, Nicolas; Mirzabaev, Alisher; Nkonya Ephraim (2013). The Economics of Land
Degradation.
110. Stellmacher, Till (2013). Local forest governance in Ethiopia: Between legal pluralism and livelihood realities.
111. Evers, Hans-Dieter; Purwaningrum, Farah (2013). Japanese Automobile Conglomerates in Indonesia:
Knowledge Transfer within an Industrial Cluster in the Jakarta Metropolitan Area.
112. Waibel, Gabi; Benedikter, Simon (2013). The formation water user groups in a nexus of central directives and
local administration in the Mekong Delta, Vietnam.
113. Ayaribilla Akudugu, Jonas; Laube, Wolfram (2013). Implementing Local Economic Development in Ghana:
Multiple Actors and Rationalities.
114. Malek, Mohammad Abdul; Hossain, Md. Amzad; Saha, Ratnajit; Gatzweiler, Franz W. (2013). Mapping
marginality hotspots and agricultural potentials in Bangladesh.
115. Siriwardane, Rapti; Winands, Sarah (2013). Between hope and hype: Traditional knowledge(s) held by
marginal communities.
116. Nguyen, Thi Phuong Loan (2013). The Legal Framework of Vietnam’s Water Sector: Update 2013.
117. Shtaltovna, Anastasiya (2013). Knowledge gaps and rural development in Tajikistan. Agricultural advisory
services as a panacea?
118. Van Assche, Kristof; Hornidge, Anna-Katharina; Shtaltovna, Anastasiya; Boboyorov, Hafiz (2013). Epistemic
cultures, knowledge cultures and the transition of agricultural expertise. Rural development in Tajikistan,
Uzbekistan and Georgia.
119. Schädler, Manuel; Gatzweiler, Franz W. (2013). Institutional Environments for Enabling Agricultural
Technology Innovations: The role of Land Rights in Ethiopia, Ghana, India and Bangladesh.
120. Eguavoen, Irit; Schulz, Karsten; de Wit, Sara; Weisser, Florian; Müller-Mahn, Detlef (2013). Political
dimensions of climate change adaptation. Conceptual reflections and African examples.
121. Feuer, Hart Nadav; Hornidge, Anna-Katharina; Schetter, Conrad (2013). Rebuilding Knowledge. Opportunities
and risks for higher education in post-conflict regions.
122. Dörendahl, Esther I. (2013). Boundary work and water resources. Towards improved management and
research practice?
123. Baumüller, Heike (2013). Mobile Technology Trends and their Potential for Agricultural Development
124. Saravanan, V.S. (2013). “Blame it on the community, immunize the state and the international agencies.” An
assessment of water supply and sanitation programs in India.
125. Ariff, Syamimi; Evers, Hans-Dieter; Ndah, Anthony Banyouko; Purwaningrum, Farah (2014). Governing
Knowledge for Development: Knowledge Clusters in Brunei Darussalam and Malaysia.
126. Bao, Chao; Jia, Lili (2014). Residential fresh water demand in China. A panel data analysis.
127. Siriwardane, Rapti (2014). War, Migration and Modernity: The Micro-politics of the Hijab in Northeastern Sri
Lanka.
128. Kirui, Oliver Kiptoo; Mirzabaev, Alisher (2014). Economics of Land Degradation in Eastern Africa.
129. Evers, Hans-Dieter (2014). Governing Maritime Space: The South China Sea as a Mediterranean Cultural Area.
130. Saravanan, V. S.; Mavalankar, D.; Kulkarni, S.; Nussbaum, S.; Weigelt, M. (2014). Metabolized-water breeding
diseases in urban India: Socio-spatiality of water problems and health burden in Ahmedabad.
131. Zulfiqar, Ali; Mujeri, Mustafa K.; Badrun Nessa, Ahmed (2014). Extreme Poverty and Marginality in
Bangladesh: Review of Extreme Poverty Focused Innovative Programmes.
132. Schwachula, Anna; Vila Seoane, Maximiliano; Hornidge, Anna-Katharina (2014). Science, technology and
innovation in the context of development. An overview of concepts and corresponding policies
recommended by international organizations.
133. Callo-Concha, Daniel (2014). Approaches to managing disturbance and change: Resilience, vulnerability and
adaptability.
134. Mc Bain, Florence (2014). Health insurance and health environment: India’s subsidized health insurance in a
context of limited water and sanitation services.
135. Mirzabaev, Alisher; Guta, Dawit; Goedecke, Jann; Gaur, Varun; Börner, Jan; Virchow, Detlef; Denich,
Manfred; von Braun, Joachim (2014). Bioenergy, Food Security and Poverty Reduction: Mitigating tradeoffs
and promoting synergies along the Water-Energy-Food Security Nexus.
136. Iskandar, Deden Dinar; Gatzweiler, Franz (2014). An optimization model for technology adoption of
marginalized smallholders: Theoretical support for matching technological and institutional innovations.
137. Bühler, Dorothee; Grote, Ulrike; Hartje, Rebecca; Ker, Bopha; Lam, Do Truong; Nguyen, Loc Duc; Nguyen,
Trung Thanh; Tong, Kimsun (2015). Rural Livelihood Strategies in Cambodia: Evidence from a household
survey in Stung Treng.
138. Amankwah, Kwadwo; Shtaltovna, Anastasiya; Kelboro, Girma; Hornidge, Anna-Katharina (2015). A Critical
Review of the Follow-the-Innovation Approach: Stakeholder collaboration and agricultural innovation
development.
139. Wiesmann, Doris; Biesalski, Hans Konrad; von Grebmer, Klaus; Bernstein, Jill (2015). Methodological Review
and Revision of the Global Hunger Index.
140. Eguavoen, Irit; Wahren, Julia (2015). Climate change adaptation in Burkina Faso: aid dependency and
obstacles to political participation. Adaptation au changement climatique au Burkina Faso: la dépendance à
l'aide et les obstacles à la participation politique.
http://www.zef.de/workingpapers.html
ZEF Development Studies
edited by
Solvay Gerke and Hans-Dieter Evers
Center for Development Research (ZEF),
University of Bonn
Caleb R.L. Wall, Peter P. Mollinga (Eds.)
Fieldwork in Difficult Environments.
Methodology as Boundary Work in
Development Research
Vol. 7, 2008, 192 p., 19.90 EUR, br. ISBN 9783-8258-1383-3
Shahjahan H. Bhuiyan
Benefits of Social Capital. Urban Solid Waste
Management in Bangladesh
Vol. 1, 2005, 288 p., 19.90 EUR, br. ISBN 38258-8382-5
Solvay Gerke, Hans-Dieter Evers, Anna-K.
Hornidge (Eds.)
The Straits of Malacca. Knowledge and
Diversity
Vol. 8, 2008, 240 p., 29.90 EUR, br. ISBN 9783-8258-1383-3
Veronika Fuest
Demand-oriented Community Water Supply in
Ghana. Policies, Practices and Outcomes
Vol. 2, 2006, 160 p., 19.90 EUR, br. ISBN 38258-9669-2
Caleb Wall
Argorods of Western Uzbekistan. Knowledge
Control and Agriculture in Khorezm
Vol. 9, 2008, 384 p., 29.90 EUR, br. ISBN 9783-8258-1426-7
Anna-Katharina Hornidge
Knowledge Society. Vision and Social
Construction of Reality in Germany and
Singapore
Vol. 3, 2007, 200 p., 19.90 EUR, br. ISBN 9783-8258-0701-6
Irit Eguavoen
The Political Ecology of Household Water in
Northern Ghana
Vol. 10, 2008, 328 p., 34.90 EUR, br. ISBN 9783-8258-1613-1
Wolfram Laube
Changing Natural Resource Regimes in
Northern Ghana. Actors, Structures and
Institutions
Vol. 4, 2007, 392 p., 34.90 EUR, br. ISBN 9783-8258-0641-5
Lirong Liu
Wirtschaftliche Freiheit und Wachstum. Eine
international vergleichende Studie
Vol. 5, 2007, 200 p., 19.90 EUR, br. ISBN 9783-8258-0701-6
Phuc Xuan To
Forest Property in the Vietnamese Uplands. An
Ethnography of Forest Relations in Three Dao
Villages
Vol. 6, 2007, 296 p., 29.90 EUR, br. ISBN 9783-8258-0773-3
Charlotte van der Schaaf
Institutional Change and Irrigation
Management in Burkina Faso. Flowing
Structures and Concrete Struggles
Vol. 11, 2009, 344 p., 34.90 EUR, br. ISBN 9783-8258-1624-7
Nayeem Sultana
The Bangladeshi Diaspora in Peninsular
Malaysia. Organizational Structure, Survival
Strategies and Networks
Vol. 12, 2009, 368 p., 34.90 EUR, br. ISBN 9783-8258-1629-2
Peter P. Mollinga, Anjali Bhat, Saravanan V.S.
(Eds.)
When Policy Meets Reality. Political Dynamics
and the Practice of Integration in Water
Resources Management Reform
Vol. 13, 2010, 216 p., 29.90 EUR, br., ISBN
978-3-643-10672-8
Irit Eguavoen, Wolfram Laube (Eds.)
Negotiating Local Governance. Natural
Resources Management at the Interface of
Communities and the State
Vol. 14, 2010, 248 p., 29.90 EUR, br., ISBN
978-3-643-10673-5
Martha A. Awo
Marketing and Market Queens - A study of
tomato farmers in the Upper East region of
Ghana
Vol. 21, 2012, 192 S., 29.90 EUR, br., ISBN 978-3643-90234-4
William Tsuma
Gold Mining in Ghana. Actors, Alliances and
Power
Vol. 15, 2010, 256 p., 29.90 EUR, br., ISBN
978-3-643-10811-1
Asghar Tahmasebi
Pastoral Vulnerability to Socio-political and
Climate Stresses - The Shahsevan of North Iran
Vol. 22, 2013, 192 S., 29.90 EUR, br., ISBN 9783-643-90357-0
Thim Ly
Planning the Lower Mekong Basin: Social
Intervention in the Se San River
Vol. 16, 2010, 240 p., 29.90 EUR, br., ISBN
978-3-643-10834-0
Anastasiya Shtaltovna
Servicing Transformation - Agricultural Service
Organisations and Agrarian Change in PostSoviet Uzbekistan
Vol. 23, 2013, 216 S., 29.90 EUR, br., ISBN 9783-643-90358-7
Tatjana Bauer
The Challenge of Knowledge Sharing - Practices
of the Vietnamese Science Community in Ho Chi
Minh City and the Mekong Delta
Vol. 17, 2011, 304 p., 29.90 EUR, br., ISBN
978-3-643-90121-7
Hafiz Boboyorov
Collective Identities and Patronage Networks
in Southern Tajikistan
Vol. 24, 2013, 304 S., 34.90 EUR, br., ISBN 9783-643-90382-2
Pham Cong Huu
Floods and Farmers - Politics, Economics and
Environmental Impacts of Dyke Construction in
the Mekong Delta / Vietnam
Vol. 18, 2012, 200 p., 29.90 EUR, br., ISBN 978-3643-90167-5
Simon Benedikter
The Vietnamese Hydrocracy and the Mekong
Delta. Water Resources Development from
State Socialism to Bureaucratic Capitalism
Vol. 25, 2014, 330 S., 39.90 EUR, br., ISBN 9783-643-90437-9
Judith Ehlert
Beautiful Floods - Environmental Knowledge and
Agrarian Change in the Mekong Delta, Vietnam
Vol. 19, 2012, 256 S., 29,90 EUR, br, ISBN 978-3643-90195-8
Sven Genschick
Aqua-`culture´. Socio-cultural peculiarities,
practical senses, and missing sustainability in
Pangasius aquaculture in the Mekong Delta,
Vietnam.
Vol. 26, 2014, 262 S., 29.90 EUR, br., ISBN 9783-643-90485-0
Nadine Reis
Tracing and Making the State - Policy practices
and domestic water supply in the Mekong Delta,
Vietnam
Vol. 20, 2012, 272 S., 29.90 EUR, br., ISBN 978-3643-90196-5
Farah Purwaningrum
Knowledge Governance in an Industrial
Cluster. The Collaboration between AcademiaIndustry-Government in Indonesia.
Vol. 27, 2014, 296 S., 39.90 EUR, br., ISBN 9783-643-90508-6
Panagiota Kotsila
Socio-political and Cultural Determinants of
Diarrheal Disease in the Mekong Delta.
From Discourse to Incidence
Vol. 28, 2014, 376 S., 39.90 EUR, br., ISBN 9783-643-90562-8
http://www.lit-verlag.de/reihe/zef
Zentrum für Entwicklungsforschung
Center for Development Research
University of Bonn
Working Paper Series
Authors:
Doris Wiesmann, Hans Konrad Biesalski, Klaus von Grebmer, Jill Bernstein
Contacts:
d.wiesmann@gmx.com, hans-k.biesalski@uni-hohenheim.de,
k.vongrebmer@cgiar.org, jtwbernstein@yahoo.com
Photo:
Welthungerhilfe/Silke Wernet, Banashyam Nagar, West Bengal, India, 2011
Published by:
Zentrum für Entwicklungsforschung (ZEF)
Center for Development Research
Walter-Flex-Straße 3
D – 53113 Bonn
Germany
Phone: +49-228-73-1861
Fax: +49-228-73-1869
E-Mail: zef@uni-bonn.de
www.zef.de